Linux Apps On Solaris

querencia writes "Sun has announced that Solaris 10 will comply with the Linux Standard Base specification, thus allowing Linux apps to run unchanged on Solaris. This isn't emulation -- they claim that it is 'kernel-integrated and supported as an operating system feature.' While I appreciate the benefits of the Solaris OS, I've considered them on the losing end of the battle until now. Will the power of Linux apps put Solaris back into the running?" Update: 08/04 15:50 GMT by J : At OSCON, Sun reaffirmed that Solaris 10 will be open-sourced. They said it would be one of the OSI licenses, not sure which yet; that this was approved at the highest levels of the company; and (with the expected "we're just guessing" language), it could happen as soon as year's end.

Comments

you mean like...

[Anonymous Coward]

you mean like how wine put linux back in the running by allowing it to run windows apps? (or at least games...)

by the way, how well does doom 3 run under wine?

Re:you mean like...

[vuvewux]

The problem is that this does the opposite - WINE takes Linux out of the running because there is now less of an incentive to write OPEN applications. The Doom 3 Linux port should be out soon if I have my way.

Re:you mean like...

[Nailer]

by the way, how well does doom 3 run under wine?

Dunno, but I hear Far Cry works fine.

And Wine does more than games. BTW - Office XP, Flash / Dreamwaver MX, iTunes, etc.

Re:you mean like...

[zz99]

The Doom 3 Linux port should be out soon if I have my way.

According to the .plan of the ID software CEO there will be a Linux version soon:

Mac and Linux: Unfortunately I don't have dates for either of these. However, Linux binaries will be available very soon after the PC game hits store shelves. There are no plans for boxed Linux games. More remains to be done for the OSX version of DOOM 3 and that will take some time. We won't release the OSX version until it's just as polished as the PC version. The date for OSX DOOM 3 remains "when it's done", but I can confirm that it's definitely coming.

Re:you mean like...

[Anonymous Coward]

Uh, you do realize that linux binaries will be officially released by id software in just a few days?

Oh, and wine is an implementation of a part of the Windows API on Linux. That's completely different from what this mentions. The OS actually operates in such a way that it will just work.

Re:you mean like...

[Nailer]

by the way, how well does doom 3 run under wine?

*Does 5 minutes of research*

Fine. You'll need this patch for Winex though.

Re:you mean like...

[mikael]

If you really want developers to consider Linux as a viable to Windows, then the first thing that should be done is to make it easy to port projects designed for MFC/Visual Studio into KDdevelop projects.

Re:you mean like...

[LittleBigLui]

by the way, how well does doom 3 run under wine?

On my machine, as good as under windows directly. See here.

Re:you mean like...

[Tony Hoyle]

If you really want developers to consider Linux as a viable to Windows, then the first thing that should be done is to make it easy to port projects designed for MFC/Visual Studio into KDdevelop projects.

Uhh Why?

MFC is dead. It hasn't been updated since version 4.2 and even MS don't use it any more. It was a hideous abortion to start with anyway.

Porting a project that's written properly is trivial - I write multiplatform stuff all the time . If your code is correctly written then it'll run anywhere with minor tweaks (it's no coincidence that doom3 can reach Linux so fast... it's because ID have programmers that are actually half decent).

Re:you mean like...

[KarmaMB84]

"Sun has announced that Solaris 10 will comply with the Linux Standard Base specification"

By following a speification apparently.

Re:you mean like...

[katorga]

It won't make any difference for Sun. They're finished and heading down the SGI path.

Re:you mean like...

[msh104]

this is all old news.... this was all anounced on slashdot 700 times

Re:you mean like...

[Fred_A]

How about an AMD64 linux binary (hint hint) ?

Just so that the hours (days actually, I didn't expect it to be that hard to find a distribution for AMD64 that works) I spent putting this damn box together (don't but a chaintech motherboard) weren't for nothing...

Re:you mean like...

[deaddrunk]

Uh because lots of things are written in it and making life difficult by making people rewrite their apps isn't exactly gonna speed up the adoption of Linux. I agree that in a perfect world apps writers would create decent, easily portable code, but having worked in far too many cruddy IT departments, most programmers don't have time to worry about whether their business apps are portable.

Re:you mean like...

[Urkin]

Uh I guess you don't program for a living or you would know about concept of 'legacy' development. Do you have any idea how many MFC projects are in use and actively being developed in most organizations today? Its a big potential audience.

Re:you mean like...

[sp0rk173]

You found one, right? If not, i've been using gentoo without a hitch for about 3 weeks now, and all the x86 binary FPS games, that i've tried, in portage work great (As long as you remember to use the 32 bit libs, not the 64 bit libs...since you have both on your system). Which distro are you using? Just out of curiosity..

Re:you mean like...

[tonyr60]

Ahhh, so Sun is out of business now (as in finished) and is then going to be restarted as a small niche company that appears to be able to have a future? Obviously you have access to some very closely held information.

Re:you mean like...

[mikael]

Uhh Why?

MFC is dead. It hasn't been updated since version 4.2 and even MS don't use it any more. It was a hideous abortion to start with anyway.

I will agree with you there. However, as it happens, I am porting a game engine + level editor from MFC to Linux. Porting the core geometry libraries has been easy. This only leaves the GUI interface to be ported. If there is any way possible, I'd rather not have the hassle of having to create/edit 20+ dialog menus + callback functions, then have to cut and paste the various function calls into the framework. Creating a single view window for the actual game is simple using SDL and OpenGL.

I believe ID avoid this problem by having the game engine and level editor combined together.

Re:you mean like...

[essreenim]

I believe ID avoid this problem by having the game engine and level editor combined together.
Wouldb't it be great if all games were written in Java 3d OpenGL and still really fast!
Then everything would already be multiplatform by default ..alas ; |
There's even a Linux java 3d opengl implementation - www.blackdown.org You probably know of it.

Re:you mean like...

[Fred_A]

Yes, I settled on Gentoo as well after trying MDK 10.0 pre1 (only one freely available to basic club members), Debian sid, MDK 10.0 final 32 bit version (I was getting kind of desperate)...

And finally Gentoo works fine even though I have to confess that I find the compiling a bit tedious. Although I expect that once the system is up to speed, it's not that bad... And the machine is quite fast after all.

I just wish XP would install on the damn thing so I could play a game every now and then. I've always kept a small windows partition on my main workstation and now that I finally have a powerful machine, it won't run the gaming OS :(

Re:you mean like...

[sp0rk173]

The trick to gentoo is to keep it up to date, so emerge sync and emerge -u world EVERYDAY. At the most you'll compile about 5 - 10 things, but usually you won't be compiling anything. Updating usually takes a tiny fraction of your computing time everyday, and with an x86-64 chip you won't even notice it. Also, for the games, i run the following very nicely in gentoo:

Wolfenstien Enemy Territory
America's Army (in all it's bullshit propagandizing glory)

If you've got an nvidia card, you should have no problem. The docs on the gentoo site will get that going for you. nvidia did release amd64 binary linux drivers, and they are in portage. Just be sure you load the 32-bit libGL.so.x (in /usr/lib32), otherwise the games will complain and die angrily. Hopefully that helps.

Re:you mean like...

[Fred_A]

I think I'll set "emerge sync && emerge -u world" as a cron job to run at night every other day...
OpenGL currently works fine in the 64 bit environment, I'm working on setting up 32 bit compatibility as well as sound, etc.

And I finally tricked XP into installing. So all is well. :)

No

[IceFox]

Just like with MS and OS/2 people will now make apps for Linux that oh yah work on Solaris not the other way around. As a developer it is a pretty easy choice to make and as we all know it is all about developers developers developers...

-Benjamin Meyer

Much like the way Wine works

[isolation]

You can think of this support for Linux apps on Solaris as the same way Wine works. It provides a layer of support by implementing the needed APIs without having to deal with a total emulation enviroment.

Re:Much like the way Wine works

[mr_z_beeblebrox]

You can think of this support for Linux apps on Solaris as the same way Wine works. It provides a layer of support by implementing the needed APIs without having to deal with a total emulation enviroment.

Score 1: informative? No you can not think of that as the way Wine works. The technical explanation was given they are complying with the LSB which is much like the POSIX. This is an inherent change to the Solaris Kernel not just an emulator or a set of libraries.

Re:Much like the way Wine works

[isolation]

The LSB defines a set of APIs and libraries along with the locations in the filesystem. This project adds a layer to intercept the Linux Syscalls and either redirect them or implement them as Solaris Native. This is the same thing the Wine does except that Wine exists only in userspace.

A better example would be Linux emulation on FreeBSD. Solaris is doing the same thing the FreeBSD people have been doing for years.

Interesting, but what about the other way round?

[hcdejong]

There's at least one Solaris application I'd like to run on Linux: Adobe FrameMaker.

Could this be

[Druss.the.legend]

The initial move of SUN towards an OpenSource OS, or even towards a linux based business model.

Re:Could this be

[exi1ed0ne]

No, It's the "Me Too" syndrome.

Just LSB or ABI/API too?

[Anonymous Coward]

Will Solaris simply comply with the LSB in a similiar manner as they supplied SunOS BSD tools with Solaris, or will it also be capable of running Linux ELF binaries unchanged? What about Linux-specific things such as clone()? That's not something you can emulate so easily.

It seems a bit of desperate measure. There was a time when Solaris was the leading UNIX on any platform. Now Sun seem resigned to play second fiddle..

Re:Just LSB or ABI/API too?

[aphor]

Look at the FreeBSD Linux support: a kernel module and an ELF loader that support all the Linux syscalls and can decide at load time which flavor of syscall to implement. The runtime linker/loader knows to go to a certain directory tree to get Linux shared libraries, and Solaris will probably work much as the sparc 32/64 bit stuff works now.

Re:Just LSB or ABI/API too?

[Anonymous Coward]

Yes that's what I'd expect, but just saying "We'll comply with the LSB" does not automatically mean that Solaris will include Linux binary support and syscall emulation. Although it probably will.

Re:Just LSB or ABI/API too?

[JonAnderson]

Of course it will. They even say if your app doesn't run we will fix it.

Linux APIs

[Sebby]

"Will the power of Linux apps put Solaris back into the running?"

I guess it can't hurt. Apple is also rumored to be integrating Linux API to future versions of OS X to help bring developers to the Mac side.

Re:Linux APIs

[zz99]

I guess it can't hurt

My guess is that this will be very popular among those that are planning to migrate. Now they can run things in parallell before they switch the bulk of their HW/SW....

Re:Linux APIs

[leoxx]

This is all excellent news. It means developers will only have to target Linux and not Solaris or Mac OS X in order to reach the largest number of users.

I doubt it

[metalac]

It seems that Solaris is having a real hard time getting trough no matter what. With the availability of so many BSDs and Linux distros Solaris is a lone wolf in the whole story. Also I don't think that people who are currently running Linux will be very eager to just jump up and switch since all of a sudden Solaris supports Linux binaries.

Re:I doubt it

[hachete]

http://www.theinquirer.net/?article=17627

and

http://linux.slashdot.org/article.pl?sid=04/08/0 4/ 1233241&tid=163&tid=155&tid=218

If you join the dots, you might see a survival strategy if the Big Bad Wolf comes a hunting.

h.

Re:I doubt it

[Neil Watson]

You've hit the nail on the head. The days of commercial Unix are numbered. It's sad and tragic. The big commercial Unix vendors have no one to blame but themselves. Unix was powering workstations and servers when Windows was still in its 3.1 days. That was a large lead that they petered away. Instead of spending that time improving their procduct (e.g. making their tools more functional like the GNU tools have become) big Unix sat back and did very little.

• Why didn't Sun fix their tar utility to properly strip '/'?
• Why didn't Sun fix their tar utility to add on the fly compression (-j -z anyone?)?
• Why didn't Sun ever develop a useful packet filtering application instead of relying on the ipfilter whose releases can often be worse than beta quality?
• Why are there so many different bin directories that the environment never pointed to (e.g. /usr/ucb/bin)?
• Sed, Awk, and Vi all had room for improvement. Why did they do nothing?

Re:I doubt it

[njcoder]

"Why didn't Sun fix their tar utility to properly strip '/'? "

A lot of the times, sun doesn't fix stuff so that they can maintain compatability between different versions which is one of their strong selling points. If you don't need that kind of compatability you can use the GNU version.

"Why didn't Sun fix their tar utility to add on the fly compression (-j -z anyone?)?"

I wouldn't call that a "fix" that's a feature that they chose not to implement. Why put it in when people are happy to pipe the compression tools in themselves. It gives them more flexibility to choose the versions they want and it makes it easier for tar by not having to worry about those things. Each utility serves it's purpose and you can use them together. That doesn't mean they should be integrated. So I wouldn't call it broken.

"Why didn't Sun ever develop a useful packet filtering application instead of relying on the ipfilter whose releases can often be worse than beta quality?"

What about SunScreen? In Solaris 10, they're going to have Solaris IP Filter which they claim to be enterprise class. From what I've read there is some shared code between SunScreen and ipfilter. Not sure which way it goes. I read the ipfilter guy licensed code from sun but couldn't confirm it. Also, Sun's main deployment areas are corporate data centers, telco's and isp's. These people use seperate firewalls to secure all their servers. Looks like sun has been coming around to smaller deployment users since at least Solaris 9.

"Why are there so many different bin directories that the environment never pointed to (e.g. /usr/ucb/bin)?"

Again, this is for compatability reasons. /usr/bin is the Sun versions, /usr/bin is the berkley tools, /usr/local/bin is usually where the gnu tools go. One of the best things about sun is their commitment to binary compatability. You can develop on your workstation and deploy on a e25k without making any changes. You can also deploy most applications written for prior versions on new os versions. To facilitate that and still allow people to use other tools, they set up different directories. They're not pointed to because you should only point to them if you need to.

"Sed, Awk, and Vi all had room for improvement. Why did they do nothing?"

Beats me. But you can download the gnu versions of them if you need them. Those three things have never been a bother to me in any work I've done on sun servers.

Re:I doubt it

[Neil Watson]

What about SunScreen?

Have you ever tried to use SunScreen? Very unusable in my experience. You need to be local, running X to install it. Really, who runs X on a server? The initial configuration options are limited an break any remote access you may have just had. The fact that Sun is abandoning it for IPfilter indicates that don't think SunScreen is very good either.

A lot of the times, sun doesn't fix stuff so that they can maintain compatability between different versions which is one of their strong selling points.

How well has that worked for them?

While I'm still ranting:

• How about improving the pager (more)?
• How about making sure the backspace key works?

Re:I doubt it

[njcoder]

"The fact that Sun is abandoning it for IPfilter indicates that don't think SunScreen is very good either."

I've only used SunScreen to play with and I don't have any real opinions. I'm still of the mindset that packets should be filtered before they get to the host. Also just because it didn't come with ipfilter doesn't mean you couldn't put it on there afterwards. Remember, you always had the choice, everything doesn't have to come bundled. What's good is that now Sun is making improvements to IP Filter in their Solaris IP Filter.

"How well has that worked for them?"

You'd have to ask someone at sun. The type of development I do doesn't really make that an issue for me. Though I've seen reports of where it has paid off for the customers. Especially in regards to compatability between workstations and their refridgerator sized servers. One account I read said that he took the drives out of his workstation and put them in an e15k i believe and everything worked perfectly. A lot of the big apps that run on solaris cost many thousands of dollars. Solaris then becomes a good investment for customers. You start out on solaris x, you buy some massive app for 25k, when solaris x+1 comes out and has much better performance you can upgrade the os and still have the app you spent 25k for running. Just like by default the root shell is the bourne shell. scripts written a long time ago will still work in an upgrade.

It's not for people that need to be on the cutting edge. When you look at most corporate development though, they care about stability. They spend millions of dollars building their application. They want a reliable platform to run their apps. The platform doesn't make them money, it's just a place to put what does. It should get better without out requiring them to to change their code.

"How about improving the pager (more)?

I don't care, I like less better. "How about making sure the backspace key works?"

When does it not? I have some sparc boxes and one that runs solaris x86. I haven't had a problem with backspace. It might be the terminal software you're using. Or maybe it's the bourne shell. I like bash. It's been part of the supplemental software cd for a while, as have other shells. Since solaris 9 i beleive it's been installed though it may not be the default.

Re:I doubt it

[Neil Watson]

I had backspace key issues with Sol 9, x86, out of the box, while sitting at the console.

Re:I doubt it

[ZeekWatson]

I had backspace key issues with Sol 9, x86, out of the box, while sitting at the console.

This is true for HP-UX also. Plug in a HP-UX serial console to one of their "Enterprise" servers (they have no videocard or keyboard port) login and hit backspace. It doesn't work!!!

Sure you can install GNU utilities to get around this but I'll install my GNU utilities with linux, thanks!

Also, Solaris tar is busted in that paths longer than 255 characters are truncated. Solaris tar is a POS.

Re:I doubt it

[Vladimir]

You're completely right, the user experience is terrible on Solaris. A lot of their tools have bugs (sad experience of putting 80GB of data on a tape - the only tar that worked accross Sparc/OSF/Linux, was, of course, GNU tar from my Debian distro). I also don't understand their attitude towards 64bits: they hype it left and right, but when you take a look -- most of their tools I _needed_ to support large files didn't do it out of the box (tar, gzip, sh, etc.). Even more: why awk cannot handle more than 512 fields? Why regex crashes inside libc on patterns ~40KB? Why can't you have more than 255 FILE* opened ? I really cannot think of any SUN's software that I would say, "oh, cool!"; any Linux distro provides much better environment. People say kernel is more robust, I cannot comment on that: indeed it crashes not very frequently, but so do our Dells (300 days uptime so far). In addition we have 4 times more Dells each 4 times faster than Sparc.

That's very interesting how SUNW will proceed: they still have tons of money and quality people, but what are they all doing?

Re:I doubt it

[Cajal]

Regarding the max number of file descriptors, you need to set the rlim_fd_max and rlim_fd_cur settings in /etc/system.

Btw, when will Linux finally get tools like pmap, trapstat, lockstat, ppgsz, libumem, etc? I was a longtime Linux user, but now that I'm using Solaris more, I find them hard to live without.

Re:I doubt it

[0x0d0a]

Btw, when will Linux finally get tools like

I haven't used any of these, but I think that you can get similar functionality for at least some of them (keep in mind I'm going off a quick Google for a description of each Solaris utility). Not familiar with how completely a replacement each is:

pmap

Linux pmap. If, for some reason, your system doesn't have it (can't imagine why; it's part of procps on Fedora Core 2), I imagine that you can get similar data from: /proc/<pid/maps

trapstat

oprofile

lockstat

SGI lockstat

ppgsz

I don't think that the functionality (variable page sizes) that this provides can be provided on x86 hardware, though I reckon if you're running Linux on SPARC, it's a pain.

libumem

There are a *ton* of Linux malloc() replacements and memory debugging tools. I generally use valgrind, but I've used ef, debauch, memprof, and a ton of other similar tools in the past. The standard glibc() malloc provides some basic debugging features as well.

Re:I doubt it

[Vladimir]

Note that I talked about FILE, not fd (there is no problem with file descriptors).
On the other hand, FILE is a struct defined in the following way on Solaris 8 (cut from 'gcc -E' for something that includes stdio.h)

struct __FILE
{
ssize_t _cnt;
unsigned char *_ptr;
unsigned char *_base;
unsigned char _flag;
unsigned char _file;
unsigned __orientation:2;
unsigned __ionolock:1;
unsigned __filler:5;
};

The _file here has type "unsigned char". That's it, at most 256 files open at the same time. And I guess backward compatibility prevents them from changing it to int. But think about it: I work on a box with 16GB memory and 8 Sparc CPUs, compile purely 64bits code and CANNOT have more than 256 FILEs! Why these limitations from the dawn of computer era are still there?

Sorry, can't comment about other tools you mentioned, don't you think /proc & lsof & valgrind provide comparable functionality?

Re:I doubt it

[xcomm]

You are mostly right and I think the reason they did not improve the cute bitch Solaris in the way you mention makes the difference: They improve Solaris there where PR and Management tells (remember all the esoteric features they announced for Solaris 10), not where the programmers and the users think it should. Thats the real crux with proprietary software.

BTW:
>Why didn't Sun ever develop a useful packet filtering application instead of relying on the ipfilter whose releases can often be worse than beta quality?

You mention Iptables - you never heared of SunScreen shipped with Solaris 9? It's on of the best firewalls I ever saw!!!

Regards, xcomm

[SIG] I think we will see Osama captured right before the US election.

Re:I doubt it

[Cajal]

Regarding pmap, I meant something comparable to Solaris pmap, not the limited Linux pmap. On Solaris, pmap can display unresolved dynamic links, hardware page sizes, swap information, etc. And on Solaris 10, pmap will explicitely label the stack, can work on a core file,

Regarding ppgsz, I know that x86 chips do support multiple page sizes - Linux 2.6 includes so-called "hugetlbfs" support, but this works by allocating a fixed number of 2 MB pages. You can't set the page size on a per-process level, like you can on Solaris. Without the ability to do this, oprofile's TLB output isn't so useful. See this article on Sun Blueprints for how Solaris handles it.

Re:I doubt it

[Alex]

Have you ever tried to use SunScreen? Very unusable in my experience. You need to be local, running X to install it. Really, who runs X on a server? The initial configuration options are limited an break any remote access you may have just had.

Total crap - Sunscreen is hardly brilliant - but it is very easy to configure a basic configuration on the command line.

Alex

Re:I doubt it

[JonAnderson]

Thats a stdio limitation. man -s2 open.

Re:I doubt it

[JonAnderson]

(remember all the esoteric features they announced for Solaris 10) Name ONE esoteric feature of Solaris 10!!! I honestly can't live without Dtrace now - it TRULY is a revolution in how you approach and debug complex issues.

Re:I doubt it

[metamatic]

Plus, who wants to run SYSV when the intellectual property is controlled by a rabid SCO?

Re:I doubt it

[Vladimir]

Given that fdopen() isn't working on fd > 255 (at least on Sol8), it means any program/library that may encounter many file descriptors cannot use stdio. So how do you do thread safe I/O? How do you simulate fflush, vfprintf, etc? Write your own libc?

So what has Solaris got?

[Chemisor]

So what does Solaris have that Linux doesn't, except for the hefty price tag? It sure isn't multiprocessing anymore.

Re:So what has Solaris got?

[chegosaurus]

dtrace, zones, zfs, Sun support, source compatibility with Solaris SPARC, better stability (IMHO), and some people just prefer it. And it's not very expenive, if you pay at all.

Re:So what has Solaris got?

[njcoder]

Actually, solaris seems to be cheaper than the Enterprise branded linux distributions. In some cases WAY cheaper.

Re:So what has Solaris got?

[mslinux]

Hardware that CIO's and Lawyers tend to like.

Re:So what has Solaris got?

[linsys]

What hefty price tag are you talking about??

Soalris 10:
$99 (One-year subscription) - Commercial Use
FREE - NON Commercial

Soalris 9: New Sun Computer Systems. The end user is authorized to use the latest version of the Solaris Operating System (or any other version still commercially offered by Sun) with the new Sun computer system and system board purchased from Sun or an authorized reseller."

And if it's for development, or educational use it's FREE as well.
"

Re:So what has Solaris got?

[ArbitraryConstant]

How about a stable branch? Solaris actually has one.

The only up to date versions of Linux that can touch Solaris in scalability terms are now development versions. It's up to the distros to figure out how to make it stable.

Re:So what has Solaris got?

[stor]

Solaris is still a more stable platform that responds to load very well.

If I was asked what OS to run Oracle on in a large enterprise where rock-solid stability under load is the number one criteria, such as in a financial institution, I'd feel safer with Solaris but wouldn't see Linux as a particularly dangerous choice.

Solaris has had superior (in terms of stability) LVM and VM for instance. This stuff can be important in certain situations.

I have been very impressed thus far with Linux 2.6: it's the most stable and "polished" Linux Kernel series I've experienced. I haven't thrown it in production yet but plan to roll it out on a couple of the least business-critical machines in a few months time.

I think it's inevitable that Linux will surpass Solaris (and all other Operating Systems, for that matter) in almost every way but it's not there yet: Linux has evolved at a fast pace and often features have been merged that didn't turn out well at all, requiring band-aids, re-writes, bug fixes, etc. and causing unknown bugs, regressions and unmaintained code. This seems to have slowed down a lot though. Maybe it's just me but some of the Kernel devs seem a lot more quality-focused and critical now. Praise Andrew Morton.

If over the next year 2.6 keeps impressing me with it's stability, performance and responsiveness under load when I place it in production we could have a winner. Big time.

Cheers
Stor

Re:So what has Solaris got?

[mpieters]

A no-litigation contract with Microsoft for the next 10 years? Look out for the safe-haven FUD to come from SUN Real Soon Now (TM).

<tinfoilhat>
Remember how both Microsoft and Sun took out a hefty license with SCO? Looks like Sun figured out a strategy they hope will get them customers; hype the Linux Patent and Copyright-infringement FUD and provide a 'painless' migration path for all those CIOs running around like scared headless chickens.
</tinfoilhat>

Re:So what has Solaris got?

[pajs]

Thats certainly not my experience. Dell's service is awful in comparision to sun. We paid for gold sun support (4 hour call out), and a dell's 4 hour contract. Sun turned up with a replacement CPU within 3 hours when the sun system failed. Dell took 2 weeks to even accept that the motherboard was dead. And then took another 2 days before an engineer even turned up. They refused to send an engineeer to diagnose the problem. They also refused our own diagnosis.

(For those intrested, the machine locked solid whenever you tried to use the onboard network card. Dell gave us a program to run on the server that found no faults, but this program didn't even test the networking.)

Re:So what has Solaris got?

[cbreaker]

We have 4,200 users where I work. If we were to move to Solaris or Linux instead of Windows for the desktops, the $100 a seat adds up very quickly, when you have to pay every year.

Re:So what has Solaris got?

[Cajal]

Funny, I have the opposite impression of Linux 2.6. The lack of a proper development tree is pretty disturbing. And there are way too many patches going into an allegedly "stable" tree.

Re:So what has Solaris got?

[linsys]

I'm sorry but let's do some numbers because I am missing the boat:

MS Environment:
MS XP 2002 - $270.99 x 4,200 PCs = $1,138,158
Norton Anti-Virus 2004 - $36.99 x 4,200 = $155,358
Norton Internet Security - $49.99 x 4,200 = $209,958
Support - $39 x 1 = $39.00 (I figured 2 support tickets web based)

Solaris 10:
Retail price - $99 x 4,200 PCs = $415,800
AntiVirus - None Needed
Internet Security - It's called ipf and it's free
Support - FREE 90 Days

TOTALS:
MS = $1,503,513
Solaris = $415,800

Now this doesn't take into account group discounts, I know volume discounts are available for all products but still we are looking at 1/3 the cost to implement Solaris.

This also doesn't take into account problems like CodeRed, Slapper, or Blaster. I know I was NEVER affected by any of these using Solaris or Linux and my NT team was hammered by calls and problems from inside the company.

This doesn't take into account uptime vs. downtime. I know from EXPERIENCE that Solaris Servers have a much higher availability then Windows Servers, how much money is lost due to downtime is something which has to be handled on a customer by customer basis. We had a billing environment which cost us $1,600 a minute every time it was down.

There is NO comparison here...

Re:So what has Solaris got?

[sp0rk173]

Evil ties aside - is this a bad thing? It gets linux into the enterprise environment, providing CIO's the piece of mind they need to roll out a whole new production environment. Granted, in the sort term it could take market share away from Red Hat and Novel/SuSE, but in the long term if those two companies provide a support solution that is proven, for cheap, and all this SCO garbage quiets down...which it will..they stand to profit. This could be a very good thing for the future of linux.

Re:So what has Solaris got?

[guacamole]

You can get FREE security updates and many bugfixes for something like six years. In the process, they also keep updating the OS installer so that it could be installed on hardware released after the original OS release. Show me a Linux distro that does that either for free or for a nominal ONE TIME fee (when you buy the OS). It is a big thing for organizations that want to standardize on one OS release without having to upgrade and validate all of their applications with the new OS every year.

Re:So what has Solaris got?

[black mariah]

Are you Icelandic or are you retarded?

Solaris 10: Retail price - $99 x 4,200 PCs = $415,800 x THE NUMBER OF YEARS YOU BUY SUBCRIPTIONS

That's $99 PER YEAR, PER LICENSE. That's, by your calculations, $415,800 out of pocket EVERY FUCKING YEAR. Also, it should be incredibly fucking obvious that someone with 4200 installs to worry about isn't going to be settling for the fucking 90 day free support.

You're right, there's no comparison, but it's because you're a fucking idiot.

Re:So what has Solaris got?

[cbreaker]

Phew. I couldn't have said it better myself.

And like we'd run "Norton Internet Security" or NAV2004 on nearly 5,500 PC's. (I said users, not PC's. Many people have multiple computers + notebooks and such.)

I mean, I couldn't get the company to switch away from MS unless I was *much* higher up in IT anyways, but to say $99 isn't a big deal is ridiculous. It's subscription based service, and we could use that extra half a million on support with Linux - added with the existing support budget we'd get friggin' gold carpet treatment, not 90 day free end user support.

Well, I guess he's never worked in a large enviornment before; pretty obvious if he thinks that anyone without more then a couple thousand machines isn't going to have a Premiere support contract with MS anyways. (big bucks.) Plus, his figures are way low; it's much more expensive then that to have a full Microsoft shop when you add everything together.

Linux/OSS would be a lot cheaper in the long run for us, especially because we have a full development staff in house developing all our own applications. Many of them already run on IBM UNIX servers, too, so it wouldn't be extremely difficult to get some code changed to be Linux desktop friendly.

But, some people want to just keep things they way they are even if it's more money. We've got a full Windows IT group and desktop support department so it would mean a huge transition in personel and a lot of end user training. It would be painful for the company, so I guess the money is worth it in their minds..

It might not always be this way though, so there's always hope for the future.

Re:So what has Solaris got?

[CaptainAvatar]

Very good points. OTOH, if you figure a 3 or 4 year rolling upgrade cycle for the Windows machines then over the cycle the costs will be similar on the OP's figures - still not exactly a big argument for moving to Solaris. And of course hardware is more expensive for the Suns, as well.

Re:So what has Solaris got?

[JonAnderson]

No it isn't.

Re:So what has Solaris got?

[CaptainAvatar]

Looks like somebody's karma whoring for that +5, Informative post.

I can't wait...

[halivar]

It's just Solaris with glibc.

I can't wait for RMS to start demanding people call it GNU/Solaris.

Re:I can't wait...

[mslinux]

I can't wait for RMS to start demanding people call it GNU/Solaris.

RMS died of a massive heart attack when he discovered that GNU had been certified Unix... He kept mumbling, "How can 'GNU's not Unix' be 'Unix'," while drool ran down his chin.

Re:I can't wait...

[miniRMS]

Yes. I will do straight away. R

Re:I can't wait...

[demachina]

The burning question is how can SUN open source Solaris without further undercutting SCO's case or joining the ranks of the sued. I'm pretty sure Solaris must have encumbrances from SCO/ATT Unix. Did they get SCO to renounce all rights with the $10 million dollar payoff. Regardless if SUN open sources things like ELF I can't see how this wouldn't knock another leg out from under the teetering chair that is the only thing left keeping SCO from finishing the job of hanging themselves.

Re:I can't wait...

[dourk]

GNU's Now Unix.

Re:I can't wait...

[0x0d0a]

Actually, Unix is now GNU.

But, yes, your approach is elegant.

Really, Stallman should sit down and have a big bash (heh, heh, heh... bad pun). He wanted to provide a free *IX alternative to the world, and he and the other GNU contributors ended up taking it over.

Re:I can't wait...

[JeffTL]

Easy. Retcon it to "GNU's Now Unix"

Re:IOW...

[tekunokurato]

It's a good point, though. Linux will never have any application competitive advantage; only that of its core operating system's reliability, functionality, etc. Any application that will ever be developed for linux will be snatched up by anyone willing to implement a few relatively simple APIs in their OS.

Cheap way to develop for both?

[grunt107]

Although it seems a doomed strategy, Sun could be allowing for an internal Linux development path which they could then back-port' to Solaris, allowing Solaris to expand its portfolio.

This would, IMO, backfire since a potential customer would see Linux as the more influential and therefore desirable IT tool.

Re:Cheap way to develop for both?

[Stone316]

The thing that scares people about linux as with other opensource software is support. That is changing with the likes of Redhat, Suse, etc. But the stigma is still there... Managment feels all nice and fuzzy when they can point blame at a big company like Sun.

Re:Cheap way to develop for both?

[mewphobia]

This would, IMO, backfire since a potential customer would see Linux as the more influential and therefore desirable IT tool.

Do you think the same thing has happened with wine? People know wine exists, so they think windows is more influential and develop for it? Are we shooting ourselves in the foot by creating wine?

What are the differences in the wine situation that make it okay to support another platform's API? Just curious...

Note this is only for Solaris x86

[MojoRilla]

This only works on Solaris x86 machines, which has always been the ugly Solaris step-child.

This seems to me to be a little desperate. Sun seems to be saying that Linux has won, at least in terms of software support.

Re:Note this is only for Solaris x86

[qweqazfoo]

Ummm, why wouldn't this work with Linux sparc binaries? Sure, they're rare, but people do run Linux on old sparc hardware. Seems like it wouldn't be worth all this effort JUST to do the API port to Slowlaris x86.

Re:Note this is only for Solaris x86

[jgardner100]

Well, Solaris has actually allowed this for a long time via lxrun (all that's needed is to translate the linux system calls to Solaris, Xwindows etc remain the same) so all as they are doing is moving it into the kernel. It's a logical step as far as I can see. Does Wine mean that linux lost to Windows, of course not.

Re:Note this is only for Solaris x86

[BJH]

The reason it wouldn't work with Linux SPARC binaries is that Sun isn't going to be supporting that, which you'd know if you'd read the article.

Re:Note this is only for Solaris x86

[civilizedINTENSITY]

Makes me wonder about that "nominal 5%" performance penalty. Thats 5% slower than a binary that wouldn't require the translation, yes? But as I recall from earlier Solaris x86 postings slashdotters considered Solaris x86 a dog. So the real question is how much slower will linux binaries run on solaris than on linux? Has any optimizaton gone into Soalris x86? Or is it still a dog?

Re:Note this is only for Solaris x86

[illumin8]

This only works on Solaris x86 machines, which has always been the ugly Solaris step-child.

Your point is taken, but with the release of full 64-bit Solaris 10 for X86-64 and Sun's new 2-way AMD Opteron workstations, 2 and 4-way Opteron servers, and soon to be released 8-way Opteron servers, Sun is betting the farm on X86-64. The plain and simple truth of the matter is that Opteron offers two to three times the performance of current UltraSparc chips, and I predict that Sun will replace their entire product line, except for the extreme high-end, with Opteron, in the next 5 years.

The other thing you should consider is that more Sysadmins know Solaris than any other flavor of Unix, so giving them the capability of running 64-bit Solaris with 32 or 64-bit Linux applications side-by-side is clearly a winning move on Sun's part. Now, if only they can execute properly. Some of the benchmarks on the new Java Workstations (I don't know why they call them that when they're really just AMD Opteron workstations) have them running the BLAST benchmark on Solaris 10 x86 up to 61 percent faster than a Dell Precision Workstation running Linux.

Given the choice between a 32-bit Dell Xeon workstation with no console port, running Redhat, and a real 64-bit Sun workstation with a console port and everything, running Solaris 10, with full Linux compatibility (or dual-booting to Redhat if I so desire), at a lower price, guess which one I'm going to choose?

Re:Note this is only for Solaris x86

[Anonymous Coward]

The comparison with Dell is misleading at best.

Lets make a reality check.

Sun's cash-cow has been the SPARC/Solaris combination. Lately this cash-cow has been doing less and less good and that is why Sun has been in trouble. Making profit on x86 servers is very hard, IBM and HP have very low profit on their x86 business, Dell is doing OK. There is no reason to believe that Sun will be saved with the x86 servers and will not become their next cash-cow.

Add to this the reputation that Sun has as a x86 server vendor. They have a decade of history of downtalking x86.

They need to do something else and they seem to go to a more software and services company route now. That may go OK, but who knows...

Re:Note this is only for Solaris x86

[Cajal]

Well, it's about a year old, but you might want to take a look at this story on osnews that compared RedHat 9 and Solaris 9.

Re:Note this is only for Solaris x86

[puppetluva]

I agree with most of you points except one. . more Sysadmins know Linux than Solaris at this point. Commodity hardware swamped the market with both Linux boxes and Linux admins.

Re:Note this is only for Solaris x86

[T-Ranger]

Imagine how much Sun will charge for the same IDE hard drive the corner store charges $99 for, once blesed by Sun, and put in 'format's database. I perdict it will list at $749.

Re:Note this is only for Solaris x86

[JonAnderson]

http://store.sun.com/CMTemplate/CEServlet?process= SunStore&cmdViewProduct_CP&boxid=X9217A This is just an example. Quite pulling numbers out of your ass.

The more *nix Software the better

[njcoder]

When linux first came out they had a Solaris emulation to be able to run apps made for Solaris. These days that are a lot more apps written for linux than there were back when linux first came out (not sure on the ration of software for linux vs solaris just linux then and now).

Open Source Software isn't just Linux and the GNU userland software. It covers a wide range of different software including software that runs on Linux. In the whole sea of OSS, Linux is just a one small part. This is good for OSS projects because they now have the potential for being run on a wider range of platforms without porting issues.

Solaris has always been a good operating system. You can tell the kernel devs know this as well because searching the mailing list you'll see that solaris is referenced more than any other commercial unix. There are comparisons of how the current kernel compares to the solaris kernel as well as trying to figure out how solaris does things.

Solaris 10 is going to have a lot of improvements to it as well. There are a lot of sun hardware out there and still a lot of sun hardware being sold so it helps OSS projects reach further with less work.

For the people that see open source software as only being about Linux, I don't think they'll respond as favorably.

Short Answer

[mr_z_beeblebrox]

Will the power of Linux apps put Solaris back into the running?

No.

The long answer, Linux adaptation is slow because the FUD says that Linux is too hard, so IT managers avoid it. Linux is only now gaining ground as linux devotees have waged a constant war against that FUD. The FUD sources also say that Sun is too expensive and only caters to those who can afford their proprietary hardware. Sun has not yet begun to fight the PR campaign which it will take to overcome that. My thought is that by the time Sun gains that acceptance Linux will have near equal penetration into the corporate environment as MS.

There was a beta version once !

[tarkin]

See: here

I remember running it a couple of years ago, it was just another ugly motif applications to me ;-) But I was not a pro user.

I guess the demand for publishing under Linux stalled...

Re:There was a beta version once !

[zz99]

Until Adobe started killing Framemaker, it was widely used as a word processor in multi os environments.

It didn't look great. It wasn't the most intuitive program. It wasn't the fastest. But it worked on all platforms, and the documents could be opened, edited, printed and saved on all platforms.

We put all our project documentation in it, due to our various OSes.

Now that Adobe no longer seems interested in supporting multiple platforms, we are migrating away.

Re:There was a beta version once !

[hcdejong]

There was a demand, just not at the price point Adobe wanted (FrameMaker is about $800 for the Windows version, $ 1400 for Solaris). IIRC the user survey showed Linux users expected it to be free.
It's a pity, FrameMaker still is one of the best tools around (and the most accessible) for long-document publishing, even if Adobe has been neglecting it forever.

Re:There was a beta version once !

[the_weasel]

Now that Adobe no longer seems interested in supporting multiple platforms, we are migrating away.

To what?

Re:There was a beta version once !

[Panthras]

If you like FrameMaker then take a look at KWord and you will find like-minded people. From their web site:

KWord is a FrameMaker-like word-processing and desktop publishing application

They could use some help, but the recent decision to switch to the OpenOffice file format should avoid duplication of effort in the filters department at least.

Re:There was a beta version once !

[T-Ranger]

ASCII text.

Community Software (blastwave.org)

[sudohnim]

You've never heard of CSW?

What is blastwave.org?
blastwave.org is a collective effort to create a set of binary packages of free software, that can be automatically installed to a Solaris computer (sparc or x86 based) over the network.

We (CSW) don't provide "Linux apps", but we natively compile and package software for Solaris.

Will the power of Linux apps put Solaris back into the running?

The power of free software compiled natively for my SPARC has returned Solaris to being my primary desktop. (Now if only I could afford a Blade 2500....)

Re:Community Software (blastwave.org)

[illumin8]

We (CSW) don't provide "Linux apps", but we natively compile and package software for Solaris.

A big thank you is in order to you and the other folks at CSW that have made life as a Solaris sysadmin about 100 times easier.

(Now if only I could afford a Blade 2500....)

Forget about that crufty old 2500... when you get some cash, hook yourself up with the new and improved hotness...... I'm wishing I could afford one of these right now.

Re:Community Software (blastwave.org)

[sunwukong]

If you're a US resident, try the Sun auctions.

News of the Weird

[Onimaru]

I find myself wondering what Sun's strategy is. I mean, they go to battle with MS, enter a closed room, and come out best buds. Then they rail against FOSS in favor of open standards and threaten to do a hostile takeover on a leading Linux company. So then you think they've gotten a big check and become a patsy, right?

And throughout this blustering, they put forward the idea that through buying Novell they can somehow "own" the OS IBM is married to, which is kind of missing the point of Linux, but right in line with SCO's claims

Then they come out with news like this. As far as I can tell, their reasoning goes like this:

1. Microsoft's business practices are bad
2. But Microsoft is good
3. Open source is bad, you should run software implementing open standards instead
4. Especially since someone could buy your open source and yank it out from under you (but not your open standard)
5. But hey, if you want to run FOSS, we'll support your doing that.

Has anyone checked for schizophrenia?

Re:News of the Weird

[Troed]

Think about what you've written above, and then add this to it all.

Makes sense now, doesn't it?

Re:News of the Weird

[jgardner100]

Show me where Sun ever said open source was bad, look to Openoffice etc if you want to see their opinion on the topic.

Re:News of the Weird

[Stinking Pig]

Their strategy is to be the Unix king again. They're slowly building up resources to make themselves an amenable home for all the FOSS geeks. Then when Microsoft comes out with all guns blazing in the patent litigation war, they'll be a safe haven because of the agreements signed.

Re:News of the Weird

[Cajal]

Sun never threatened to buy Novell. It was essentially a random musing in a blog post by Schwartz that got blown way out of proportion.

Re:News of the Weird

[Onimaru]

There are a lot of good quotes from Schwartz here. Also good for your daily dose of venom is his weblog. The message is clear: Solaris' binary only, standards based distribution is open and good, while linux's open source has and can be twisted into being proprietary and bad.

I love OpenOffice, but this is starting to smack of the Real thread. You can't coast on past goodness (or badness) forever.

Re:News of the Weird

[Onimaru]

Hmmm...I was not aware of that. Amazing.

Re:News of the Weird

[Cajal]

But the ability to run Linux apps natively on x86 hardly seems to matter, when you can just download GPL'ed source and compile away with good chance of success.

Yes and no. I'm a maintainer for Blastwave, a community project which ports and packages open-source software for Solaris. Belive me, a lot of the GNU-based projects have a lot of cruft in them, and it can often be quite a pain getting them compiled on Solaris (gcc in particular is quite troublesome). But I don't think this is why Sun is pushing this technology. It's really geared towards running binary-only Linux/x86 software on Solaris. It's the same rationale behind FreeBSD's Linux ABI support, for example.

Yeah, of course! That makes sense!

[Schreckgestalt]

They are trying to buy Novell, who own SuSE. They probably want YaST on Solaris! Yuck!

Re:Yeah, of course! That makes sense!

[IANAAC]

They wound't have to buy anything to get YaST. It's already open-sourced and recently GPL'ed.

Finally

[Nailer]

Solaris can be considered a real Linux ;^)

Application/OS Security?

[akaiONE]

The thing I quite don't get a grip on here is how Sun can claim that Solaris is so much safer when it now can run Linux-applications. For years Sun have been preaching that applications they have are better and more secure. When they now comply with the LSB, wouldn't that make their OS just as "insecure" as Linux supposedly are in their views?

Their webpage says:
"You can safely run Solaris and Linux applications side by side in the same container, or you can configure separate containers that isolate Solaris and Linux applications from each other and from system faults. If an application fault occurs and the application needs to be restarted, other applications continue to run without interruption. ".

Okay, let's look at this. You can now run Solaris and Linux-applications side by side - This would mean a security breach in their previous views then? Or, you can choose to lock the Linux-applications away in their own container - This seem much more in line with previous statements from Sun.

"Unlike technology previously available for running Linux in other non-Linux environments, Project Janus functionality is kernel-integrated and supported as an operating system feature."

So, this LSB-compliance are kernel-integrated, and yet they claim Solaris is more secure than Linux? Can someone please help me out on this? I'll try to investigate myself, but I am not sure what I will find, as Solaris for now, still are, closed source.

Re:Application/OS Security?

[Stinking Pig]

ohmigod, marketing said something nonsensical... stop the presses!

There's a reason why they're going to be first against the wall when the revolution comes, and it's doublespeak. Marketing blathers on about what they think you want to hear now. Remembering what they said yesterday is a pointless exercise, they were lying then just as much as they're lying now.

I may be a bit bitter, I'm just about to present my company's product in a marketing event where one of the other vendors is going to tell people that they don't have multi-tasking unless they're using a P4 with HT. Oh yes, and security updates won't work if you don't have multi-tasking.

Re:Application/OS Security?

[pwags]

How could the mere ability to run application "X" make an OS less secure?

Anyone can take a secure OS and by doing things to it make it less secure. If someone puts an application on that runs will root privledges with a security hole in it that allows for remote access with root priv's, how's that the fault of the OS?

If I'm going to drive 100mph through downtown without a seatbelt in my Volvo, I'm not really going to be able to claim that a Volvo isn't safe when I get into an accident. (Though I could sue and win a legal claim with that arguement, but we're talking about reality here.)

Re:Application/OS Security?

[k98sven]

OPENSOURCE does not mean GNU.
Right.

Proprietary does not mean non-GNU.
Right.

Proprietary does not mean Commercial Software.
Wrong. "Proprietary software" = commercial software, open source or not, gnu or not, whatever.

Open-source means I can view the source!
Wrong. Open source as most define it, by the OSI definition means far more than just viewing the source.

Proprietary means the product does not follow open standards.

Wrong again. You are confusing "proprietary standards" which mean standards which are owned by some corporate entity, with proprietary software, which is software owned by some corporate entity. Proprietary software may use a proprietary standard, but there is nothing stopping proprietary software from using open standards. HTML is an open standard, Microsoft Word format a proprietary one.

Re:IOW...

[Anonymous Coward]

That's right. However, why would this be a bad thing? The whole reason I moved to Linux was to avoid lock-in. If they port apps from Linux to Windows/Solaris/etc, yes, it's one less reason to move to Linux but it's also one less reason to stay on Windows/Solaris/etc.

Which apps, exactly?

[YellowBook]

Will the power of Linux apps put Solaris back into the running?

Which apps would those be, exactly? Just about everything significant that's available for Linux is available as source, and most of those build with autoconf and GNU tools for portability, so installation on Solaris is just a 'configure; make; make install' away.

There are a handful of proprietary applications for Linux that might be relevant, but I'd guess most of these are back-office type things that probably already have Solaris versions. That just leaves things like the Flash plugin, and I simply can't see that sort of thing as being very important.

Re:Which apps, exactly?

[JamesKPolk]

You'd be amazed at how much non-portable garbage GNU-using developers cram into applications. Gratuitous GNUisms all over the place...

Just browse through the patches in a BSD ports collection sometime if you want to see what I mean.

Re:Which apps, exactly?

[chegosaurus]

Absolutely. A couple of years ago Sun made a big song and dance about lxrun, which I guess is somewhere at the base of this thing, and in all my travels as a Solaris consultant I never happened across *anyone* who had used it.

I think maybe linux has a more up-to-date Acrobat reader than Solaris, so I might use it for that. Nothing else springs to mind though.

BTW,there's already a port of the flash plugin for Sol x86, and it works just fine.

Re:Which apps, exactly?

[Nailer]

Which apps would those be, exactly? Just about everything significant that's available for Linux is available as source, and most of those build with autoconf and GNU tools for portability, so installation on Solaris is just a 'configure; make; make install' away.

A lot of Linux apps aren't portable to proprietary or non GNU Unix - they depend on glibc, for example. You could change that, but this takes away the need for that effort.

Re:Which apps, exactly?

[TwistedSquare]

GNU-using

GNUsing?

Re:Which apps, exactly?

[upsidedown_duck]

The ultimage gnutastic gnuventure: compiling GNUCash under Solaris. Not only is GNUCash a GNOME app, it's a GNOME 1.4 app, and libtool just barfs all over the place with doubly-listed libraries and unfound libraries. Bleh. There's a reason why pre-compiled GNUCash versions for Solaris seem to be stuck at 1.6. I did finally manage to get version 1.6.x compiled, but even then the graphing features segfaulted.

Re:Which apps, exactly?

[Per Abrahamsen]

Mathematica. Splus. DB2. Oracle. There are lots of proprietary server, engeneering or vertical market applications. Sun used to be the leader of these markets, but linux have been making constant inroads. It is possible that Linux now has taken the leading role.

Re:Which apps, exactly?

[jekewa]

Oracle, DB2, Splus, and Mathematica all run on Solaris. In fact, as I understand it, Oracle is developed first on Solaris and then ported to the other OSs they support.

Verticle applications will be a different story, but there's very little that can run on LINUX that can't run on Solaris already. Never mind what gets written in Java.

Re:Which apps, exactly?

[Per Abrahamsen]

I know they run on Solaris. The point adressed was that there are plenty of signficant non-free software for Linux. Oracle used to be Sun-centric as you describe, but I believe they have switched to being Linux centric. Not that it matters, the "big name" packages will obviously be available for both.

The fact that Sun is providing Linux emulation does indicate that there either already are, or that Sun estimate there will soon be, a significant amount of Linux specific software amoung the vertical applications.

OS/2 and Unixware anyone?

[M1FCJ]

Oh well, it didn't help OS/2 around 1995, it didn't hel Unixware around 2000. Why would such a move help Solaris in 2004/5? People never learn from other's mistakes and have to experience failure themselves all the time.

If you want your applications to run anywhere, use something truly portable. Java? PHP? Perl? ANSI C? Yes...

Re:OS/2 and Unixware anyone?

[Anonymous Coward]

Unfortunately, while Java is powerful there are still many lower level things that you cannot do in Java. A primary example of this is low level networking protcols. Something as simple as ping cannot be written in Java because it does not implement ICMP (ok except for UDP error messages). Recompling C code isn't really the answer to portability either and we all know that many things do not always recompile due to differences in OS implementation (ex. BPF in BSD vs. raw socket in linux). So if you're an admin without much time to spare binary compatiblity is really nice. If it was easy to just recompile or use Java or Perl then why does BSD have a linux compatability?

Darl Will Sue

[Perl-Pusher]

Isn't this just a "Linux Personality Kit" for Solaris? Is Sun infringing on SCO's IP? I can hear attack dog Darl growling in the distance. And the voice of his master Bill Gates saying 'Down Boy! We already own them!'

Re:Darl Will Sue

[metamatic]

You got modded +4 Funny, but there's a serious point here: Solaris is entirely dependent on UNIX System V intellectual property, which is the code at the heart of the SCO lawsuit. SCO could pull Sun's right to distribute Solaris, and they'd have a very strong case, far stronger than their case against Linux.

Given that we know SCO is rabid and will do anything to try and extort money, I think you'd have to be insane to switch from Linux to Solaris and put yourself at SCO's mercy.

A slower death?

[DCheesi]

It'll probably allow businesses to keep using their old Sun hardware a bit longer; they won't necessarily have to junk their Solaris boxes once they standardize on Linux for their core apps. However, I don't see it selling any new Sun product. "Oh boy, now I can pay thousands for Sun/Solaris HW/SW, so I can run the same apps I could have run on a $500 PC! Yay!!" :-)

Re:A slower death?

[Mr. Roadkill]

It'll probably allow businesses to keep using their old Sun hardware a bit longer; they won't necessarily have to junk their Solaris boxes once they standardize on Linux for their core apps.

Yay! Just like we keep using our old Sun sparc kit and our DEC Alpha gear with Debian. And our new-ish Sun Intel gear with Debian. And our IBM equipment with Debian. And that Shiny! new quad Opteron we've got on evaluation from HP on Debian. Yes, I'll really be tempted to install Solaris 10 on our Intel gear instead...

However, I don't see it selling any new Sun product. "Oh boy, now I can pay thousands for Sun/Solaris HW/SW, so I can run the same apps I could have run on a $500 PC! Yay!!" :-)

Okay, I might not like Solaris terribly much, but they do some very nice equipment that's more appropriate in server rooms than the average $500 PC. If your apps won't run noticeably better on one of their sparc systems than on bargain-basement PC gear, they're probably not that interested in getting your business anyway.

All Propriety Solutions Welcome?

[Anonymous Coward]

Will the power of Linux apps put Solaris back into the running?

No.

What does Slashdot have with this fallacy about something saving Sun? Sun's hardware is expensive - why should I buy another piece of proprietary hardware? Sun's OS isn't GPL'd (insert your favorite license) - why should I buy yet another piece of proprietary Software? Some say Sun has Java - yet another piece of proprietary software. No Sun has to compete in the open market - sink or swim.

Re:All Propriety Solutions Welcome?

[kaffiene]

Apple has a proprietary hardware monopoly for their OS and it seems to be popular. Just because you have relgious objections to proprietary hardware doesn't mean the rest of the world thinks the same way.

For the curious

[Sebby]

I tried to find the original article mentioning it but could only find this, which indicates it was originally mentioned on MacOSRumors (wow! they're actually back!).

Why do this?

[hal2814]

Why not just start pushing a Sun Linux distro instead? I heard that was in the works, but I haven't heard anything about it in quite some time. I would use Sun Linux long before using x86 Solaris w/Linux compatibility.

Re:Why do this?

[christophersaul]

Why not start pushing a Sun Linux distro instead?

There's no room for yet another distro - customers don't want it and ISVs don't want to certify against anything beyond Red Hat and Suse.

When Sun launched their Intel based server, the LX50, it came with 'Sun Linux', which was basically Red Hat 7.3. The plan was to make it a bonafide distribution, supported, developed, etc.

Customers weren't interested. They either had a distro they were used to and preferred, or they had apps that needed to run on a certain distro for them to be supported - think Oracle. There was simply no demand for it.

The desktop is different - there's a great market out there for a polished, supported business desktop based on Linux. Sun's Java Desktop System is aimed at that market.

Re:Why do this?

[ttrafford]

JDS

Re:Apache

[chegosaurus]

How about getting someone who knows what they're doing to come in to compile it for you? Apache, PHP and all their dependencies shouldn't take more than half a day for any decent admin to build from source. And they can use Sun's great compilers (soon to be available for Linux) instead of gcc.

By the way...

[Azureflare]

How well does iTunes run under wine, I mean Crossover Office?

Yeah, and how many people are switching to linux because they can run iTunes and Microsoft office in linux now?

Re:By the way...

[Azureflare]

Hmm that's interesting, I hadn't thought about actual support of CrossoverOffice. It's on the system, but how do you support it when it's really windows in wine, with proprietary stuff done by Codeweavers to make it work better?

Well, from a support perspective, crossover office isn't really that great, considering if things go wrong with the app, there's not a whole lot you can do, and customers will probably blame you.

That's where I guess you should really try to find Linux/OSS alternatives as much as possible, and if the client absolutely demands software that does run on Crossover Office, you tell them that things may go wrong, and there's not a whole lot you can do about it. But you say you support people who use it, so maybe you know a lot more about it than me =)

That doesn't sound very good, does it. I suppose vmware is really the only solution for corporate environments that need windows software.

Vmware is a great tool, and I'm glad they've lowered their prices.

Re:By the way...

[EvilAlien]

Who needs iTunes with bittorrent and XMMS?

Re:By the way...

[aristotle-dude]

Maybe people who want a usuable interface for handling multiple playlists and having features like:
smart playlists
syncing with iPod
Album art
integrated burning and ripping of CDs.
shuffle
Party Shuffle
etc...

Why would it?

[CarrionBird]

The most prominent Linux apps are open source anyway.

nope...

[aggieben]

The power of linux is not "the power of linux apps", and so linux apps running on Solaris will not really make that big a difference for Sun.

Re:No

[mccalli]

As a developer it is a pretty easy choice to make...

Yes, it's the easy choice for these developers to make. It's not the correct one though - the correct one would be to figure out your environment and build accordingly.

For example, thanks to the wonders of "./configure ; make" I now build similar software for the three Unix environments I regularly use - SPARC Solaric, x86 Debian and OS X (PPC). Never have to worry about 'personalities', it just gets compiled and run.

It certainly is about developers, but it's about those developers becoming less sloppy and making fewer assumptions about environment. In many cases the sloppiness I refer to is entirely understandable: it was a pet project, only had to run in one environment, they only had access to x86 Lionux to test under etc.. All good arguments, but they don't really apply to the kind of applications you're likely to be running on your Solaris servers. These will be mostly custom-ordered vendor jobs, and the vendors should know better.

Cheers,
Ian
(Oh, and hi Ben - fancy running into you here. I'm the person who helped you out with your old Mac format floppies).

What about Fink?

[pjt33]

Surely (clones of) most Linux software is already available through Fink?

Re:What about Fink?

[Sebby]

Right, but there might be some stuff that isn't - not all software is open source, though I really don't know of any Linux-only software not available on other platforms too, but I guess it could happen...

Just think of it as Wine for Linux apps.

Re:What about Fink?

[iroll]

As an OSX user, I've actually run accross programs available for Windows or Linux, but not the Mac--Wolfenstein: Enemy Territory, for one--so being able to run Linux binaries would be pretty useful to me.

Re:What about Fink?

[drinkypoo]

It's not (clones of) linux software, it's ports of [open source] Unix software. That's the real reason this is dumb. In general, if you take a little effort to write portable code, anything that will run on Linux will generally port "trivially" to other Unices. A lot of this stuff has already been ported to FreeBSD from wherever it was written (which was not always Linux) or was targeted at FreeBSD in the first place and so it takes almost no work to get it to run on MacOS X.

Re:What about Fink?

[Sebby]

That's what makes having a Linux API great, the company can simply recompile for the appropriate CPU without any changes (or barely any), so they can still keep it closed and make it available to another platform for no additional cost.

Re:What about Fink?

[walt-sjc]

Easy porting misses the point. One of the HUGE advantages of Linux is that it runs just about everywhere from watches, PDA's, desktop, servers, mainframes, to massive clusters. IBM realizes this, just as they realized the same concept with Java (which is why they do more with java than Sun does.) Getting somthing written on Intel linux to run on PowerPC Linux is trivial compared to porting from Linux to OSX.

Re:What about Fink?

[walt-sjc]

While it's fairly trivial to get stuff to run on OSX, getting it to run as a native OSX app is a whole different ballgame. OpenOffice is a great example of this.

Also

[bsd4me]

I may be a little dense here, but what Linux binary-only packages or code that uses Linux-only system calls are available that you would want under Solaris?

Re:Also

[yusufg]

Skype is one such binary program which comes to my mind. Maybe CrossOver office

Re:Also

[Anonymous Coward]

its not the point - the benefit will be that you can take an allready working linux application binaries and throw them down on Solaris x86 and just run it - no porting or tidying up crappy code.

You dont waste time and resources porting the app, you can generally use the same hardware and but you do it under Solaris geting the advantages of that OS (dtrace, ZFS, zones, support etc).

I think the main benifit will be that once application providers are able to see that their app can run under Solaris and people use it, they might be interested in actually doing a Solaris port which with the Solaris source code compatibility you can then just compile for SPARC, opening up more markets.

Re:Also

[linsys]

Corssover Office Runs under Solaris.

err No?

[SQLz]

Will the power of Linux apps put Solaris back into the running?"

I venture to say...hello no. This move stinks of SCO as well. Maybe Sun will try to position itself as the only 'Legal' Linux compatable OS.

useless

[wobblie]

Of course, Sun is not talking about free software here ... it's easy enough to get any of that running on solaris.

They're talking about the software - proprietary - from vendors of theirs that are switching to linux because it's a cheaper (and better) platform for most apps. So, I really must ask, what is the point?

Solaris will - for the forseeable future - still be king on the mid to high end server end. They're talking here about workstation apps in the scientific and engineering realms which are moving wholesale to linux. So in essence Sun is saying here "you can run your linux apps on your legacy Sun workstations", and not much else. It's a nice gesture, but it is no earth shaker.

Re:useless

[Cajal]

So in essence Sun is saying here "you can run your linux apps on your legacy Sun workstations", and not much else.

No.

"Legacy Sun workstations" all use SPARC-based processors. This announcement is solely about compatibility for x86 Linux apps on Solaris x86. Sun just announced their first x86-based workstations (I'm not counting the ancient 386-based one from the '80s), which run Solaris x86, Linux and Windows. All Sun is trying to do is to make life easier on their customers who want to run Solaris x86.

Re:useless

[wobblie]

All SunAll Sun is trying to do is to make life easier on their customers who want to run Solaris x86.

Well, it's worse than I thought then.

Sun needs to stop being a publically traded company. They aren't going to make it that way anymore; they rode the .bomb wave and made some money; but the party's over.

Apple, for example, can do this because they have a zealotous fan base with plenty of disposable income. Sun's base is scientists and engineers who have no control over purchasing, and moreover don't give a hoot about what OS they use.

The best thing in my mind that Sun can do now is legitimize linux on (low-end) sparc. 98% of the work is already done. Take Debian and make a meta Sun distro out of it. Sun's biggest problem right now is their pride.

Re:useless

[twistedcubic]

C'mon Sun didn't ride the ".bomb wave". Sun was a well-established company long before that. Almost all big companies and universities had serious Sun hardware in the early 90s.

Re:useless

[Cajal]

It seems you have no perspective on things. Sun has been around a lot longer than the ".bomb wave." They made a lot of pioneering developments on Unix (NFS, PAM, etc). They're still one of the largest Unix vendors in the world. Solaris 10 is looking very nice (and has some capabilities that it will take Linux years to clone).

Sun's management realized that they're in financial trouble, and they've taken steps to correct that. They're pushing Solaris x86 hard now, and they've introduced very competitively priced Opteron-based workstations and servers. Solaris licensing is now cheaper than RedHat's in many cases. And like Apple, Sun is trying to innovate their way out of trouble, rather than become yet another Linux/x86 vendor (since that strategy worked so well for SGI).

Why should Sun push Linux on SPARC? Solaris is faster on SPARC, and Sun already supports it. Where is there any evidence that there's a large market for Linux/SPARC? Further, gcc still has issues on SPARC. Your own argument is inconsistent -- SPARC-based systems cost more than x86-based ones. They always have and likely always will. x86 has economy of scale; SPARC doesn't (and SPARC is, arguably, better designed with more fault-tolerance and management abilities). If "Sun's base is scientists and enginers who have no control over purchasing" then why would Sun push Linux/SPARC when they have Solaris/SPARC?

Re:useless

[wobblie]

Duh, I realize that Sun has been around longer than than the .bomb days - but let's face it, that's when they became a powerhouse and a national brand name.

Solaris on x86 is a complete failure. Who on earth wants to run Solaris on Intel? Ridiculous. It's for developers only, pretty much. Solaris x86 is an utter waste of time for Sun as a comany.

Solaris does NOT run "faster" on sparc than linux. Have you tried it? I didn't think so. the issue between linux/solaris on sparc is purely FEATURES (albeit kernel level features), and nothing else.

Re:useless

[Cajal]

Not really, Sun was a major workstation company in the mid and late '80s, long before the .bomb days. Sun made a lot of many during the .com craze, but they were a major player before that as well.

As for Solaris/x86, I'm sorry, but the facts just don't back you up. When Sun most recently tried to kill Solaris/x86, it was the user community who compalined to Sun to get it back. See the thread here for more info. Further, if you look at Blastwave's main mirror stats, you'll see that the x86 packages are downloaded quite a lot (granted, not as much as the sparc packages). See here and here.

As for Linux on SPARC, I have tried it. Over the years, I've had about a dozen SPARC machines, both 32- and 64-bit. Some versions of Linux were slightly faster on older 32-bit machines, but even on something as old as an Ultra 2, Solaris is a faster, especially the newer versions.

Re:useless

[jbester1]

Ugh, this is just ignorant.
There are plenty of programs that don't work at all out of the box on Solaris x86 (specifically OSS lisp implementations) which hopefully this will cover.

If they can get CMUCL to work on Solaris x86, I'll move my old fileserver over in a heartbeat.

This may be a new SUNrise.... or maybe not

[Anonymous Coward]

I had a presentation from SUN yesterday on Solaris 10.

Essentially Solaris 10 is going to be a huge change. SUN states they are aiming to be the best UNIX solution out there. With the amount of money they are spending/investing in developing Solaris 10 I believe they are making a very good attempt.

1. Linux apps will run on Solaris 10 on Intel/Sparc. Someone said this is just for X86.
2. DTrace a developer's sweetheart.
3. A new filesystem that will be much better than UFS
4. N1 Grid Containers. Making that purchase of the big iron more attractive. Equivalent to LPAR on mainframe.
5. Even better Multi-Processor efficiency. Linux is making good ground here but Solaris still is years ahead on many cpu's.
6. Of course, more efficient OS, better tcp/ip stack, security, etc. etc. The things you expect to improve with a new OS.

In my opinion, Solaris 10 if it meets what they
are marketing will prove itself. If not, watch
the SUN set.....

Not quite like....?

[LondonLawyer]

Isn't WINE an emulator?

From the Slashdot summary:

This isn't emulation -- they claim that it is 'kernel-integrated and supported as an operating system feature'.

Re:Not quite like....?

[the real darkskye]

Wine Is Not an Emulator, its a recursive acronym, like GNU

Re:Apache

[linsys]

What complexity of getting php to work??

If you can't run:

rpm -Uvh php-4.3.8-2.1.i386.rpm then it's hard?

and

rpm -Uvh apache2-2.0.47-1.7.2.i386.rpm

then it's HARD???

Try this:

1) Visit Apache's Web Site

2) Download httpd-2.0.50.tar.gz

3) Build Apache:

1. gzip -d httpd-2_0_NN.tar.gz
2. tar xvf httpd-2_0_NN.tar
3. gunzip php-NN.tar.gz
4. tar -xvf php-NN.tar
5. cd httpd-2_0_NN
6. ./configure --enable-so
7. make
8. make install

4) Visit the PHP Web Site
5) Download php-4.3.8.tar.gz

1. gtar zxvf php-4.3.8.tar.gz
2. ./configure --with-apxs2=/usr/local/apache2/bin/apxs --with-mysql
3. make
4. make install
5. cp php.ini-dist /usr/local/lib/php.ini

6) Configure httpd.conf

AddType application/x-httpd-php .php .phtml

7) Start Apache /usr/local/apache2/bin/apachectl start

Not quite, oh intelligent one.

[Ayanami Rei]

1) You need to emulate the Linux syscall interface. That means catching int 0x80 and treating it like a lcall into the kernel. It also means that when the kernel is entered via that method, that it uses a re-organized syscall table (possibly with differeing numbers of arguments and linux compatible wrappers).

2) You might need to provide linux type headers (dev_t, time_t, etc.)

3) /proc (I hope) and perhaps some /dev symlinks (I would think hd*, fd*, ttyS*...)

4) ld-linux

etc. No, it's not nearly as simple as that. But I'm impressed that glibc works on Solaris x86. I've heard of somebody being insane enough to port it sometime back, but I didn't think it'd be used for anything...

Conspiracy Theorist

[ylon]

I hate to be a conspiracy theorist, but this seems a little interesting.

First the SCO/Microsoft connections, then the Microsoft/Sun settlement... Now this? It seems odd to me that they are running in this direction in light of all of the Linux hoopla that's going around. Just look at "City of Munich Freezes Its Linux Migration" posted a little bit ago here. It almost seems like they are trying to put themselves into the position of snatching up those who are wavering on the Linux/licensing front.

Another thing

[Nailer]

Drivers. The LSB specifies the kernel too, IIRC. Linux has a metric shitload more drivers than Solaris ever will.

(PS. I'm not arguing with you. I liked your post.)

Re:Another thing

[Wesley Felter]

LSB specifies the kernel/user API. LSB does not specify the kernel/driver API (which changes all the time). Nor does LSB specify that an LSB-compliant system must support all the devices that Linux supports.

The bottom line is that Solaris doesn't/can't/won't use Linux drivers.

Errr...

[Ayanami Rei]

I think the point of the linux compatibility layer is to run COTS linux binaries, not stuff you can ./configure; make; make install. Because I think that'd be sort of dumb... don't you? Why not run native...

Great! I like it.

[kindbud]

Now I won't have to fix up all the unportable gcc/linux code you freaks like to think is cross-platform. Hah!

Re:IOW...

[oldmanmtn]

You're an idiot.

Just as a simple, grade A, introductory issue: How does a Linux application issue a system call? Using int80. How does a Solaris application issue a system call? Using syscall, sysenter, or lcall depending on the application and the version of the OS.

The two OSes don't even agree on the basic mechanism by which applications can communicate with the kernel. And you think it's just a matter of putting glibc on the CD. Put down the keyboard and go back to CS101 until you learn something.

Hmm...

[cmdrwhitewolf]

I wonder if this is the *real* reason behind their sudden interest in buying up SUSE?

I mean, did Sun recently discover that one of it's (SCO supplied) engineers misappropriated some SUSE code to make Solaris run that way? ;)

OS Diversity

[akinsgre]

The negative, among most posters, makes me wonder if OS diversity is good as long as all OSes are Linux:')

Sun has lost ground because their OS/Hardware solution is comparatively expensive; not necessarily because Solaris is not a capable OS.

It just amuses me that Windows homogeneity is bad; but Linux everywhere is good.

Re:IOW...

[scovetta]

Then why can't you just copy the glibc libraries into \winnt\system32?
</not-serious>

The "GNU" part of GNU/Linux

[shatfield]

This is why the "GNU" part of "GNU/Linux" should NOT be forgotten. People in the Microsoft mind-set immediately think that "Linux" is what they see when they look at a screenshot of X11 running KDE. The situation really sinks in when you realize that Linux is just the kernel, and they could be looking at *BSD, or even Darwin (Mac OS X's base), running X11 and KDE. Why not Solaris? Solaris is going one further though -- how about not having to recompile those apps that have been compiled to run on Linux? Very cool stuff indeed... especially if/when they open source Solaris! If they do it right (meaning - GPL compatible), then we'll see "GNU/Solaris", and Stallman will have a whole new name to complain about...

Re:The "GNU" part of GNU/Linux

[Perl-Pusher]

This is why the "GNU" part of "GNU/Linux" should NOT be forgotten. People in the Microsoft mind-set immediately think that "Linux" is what they see when they look at a screenshot of X11 running KDE

So is it GNU/KDE now? Since when did the GNU foundation claim ownership of KDE? If Stallman wants to call it GNU linux fine, but really shouldn't it be "KDE/GNOME/GNU/X.org/And a whole shitload of university students & profs/Many other bright people/Nasa ethernet drivers/NSA Security Enhancements/Linux"?

Re:The "GNU" part of GNU/Linux

[shatfield]

Your inane ranting aside, the underlying statement still rings true. When someone looks at "Linux", they don't even SEE Linux.

I once went to a drag strip and saw some cars that drag racers had modified. There were different makes of cars there -- Chevy, GM, Mopar, Ford, etc. Seeing "Linux" today reminds me of one of the more heavily modified drag cars that I saw. It was a 1976 Ford Maverick, and it had everything a righteous drag car should have -- tubbed rear end and all. The thing that made it way radical is that it was running a Chevy 350 engine under the hood. Looking at the car, you saw a Maverick. When it went down the track, you saw (the taillights, if you unlucky enough to be in the other lane) a Maverick going down the track... and only the most astute of observers would have heard the familiar roar of a Chevy engine.

What we call "Linux" today is exactly like that car. Linux is the engine, and "other software" - X.org, KDE, Gnome, etc - are what the user "sees". Linux users probably wouldn't be able to tell you that they were actually running on FreeBSD without popping up a terminal. Or Solaris, for that matter.

It isn't until you get under the hood that you see which engine is powering the car.

So the question that remains to be answered is (and in keeping with the drag racing analogy) -- when the cars go down the track, will the one running Linux beat the one running Solaris to the finnish line?

Re:The "GNU" part of GNU/Linux

[k98sven]

What terrible examples!

Neither KDE or X11 were developed within the GNU project.

Re:IOW...

[jonabbey]

Mmmm.. but the vast majority of syscalls made on a Linux system are made by glibc. They'd have to tweak the syscall interface in glibc for Solaris, but an adapted glibc would still be one of the defining features for Linux API compatibility.

Re:Apache

[NineNine]

Are you kidding? That's not complicated? The only way it could be *more* complicated and convoluted is if you were also to tell me that I need to solder something onto my motherboard to get it to work. Compare this to ASP/IIS: either it comes already installed and running, or you end up putting a CD in the drive and clicking "next" a few times.

But then, you probably wouldn't know where to start if you had to change the oil or plugs in your car...

Re:Interesting, but what about the other way round

[jmcneill]

NetBSD has the ability to run SunOS and SVR4 binaries. See the code for more details. Not sure if FrameMaker works or not, but it'd definitely be good to know!

Re:Apache

[linsys]

Please don't compare Apache/PHP/Solaris to IIS/ASP. It might not compe pre-setup out of the box like IIS/ASP, but:

a) it works

b) there isn't a new worm or virus for Apache/PHP/Solaris every week

c) Performance and scalability is better

d) You don't have to reboot Solaris every Week

e) I can't believe you compared Apache/PHP/Solaris to IIS/ASP, there is NO compareison. The ONLY way I have every been able to SECURE ISS/ASP is with an APACHE reverse PROXY SERVER!

Dude, you're getting a "troll"

[chegosaurus]

You can't go round saying things like that! Repeat after me: "Slowaris is teh sux (5 Funny)". "Who cares about Slowaris when linux does everything and its free as in beer and you get the source (5 informative)".

Now say 10 Hail Linuses, and come back when you're brainwashed.

Re:Apache

[Anonymous Coward]

And they can use Sun's great compilers (soon to be available for Linux) instead of gcc.

No, Sun is only releasing the Sun Studio 9 IDE for Linux, not the Sun compilers. SS9 will use GCC under Linux.

WHAT Linux apps?

[JessLeah]

This is not a troll; I'm a Linux user and have been since 1995, and I run Debian (so you know I'm a true blue old-skool dork, not some MS shill). But really-- WHAT apps? All the Linux apps worth running, with probably under a dozen exceptions, are either:

1) Already available for Solaris
...or...
2) Open-source and thus available for immediate porting

Come on. Think of the commercial closed-source stuff that's available for Linux, but not Solaris.

1) VMWare.
2) Uhh... VMWare.
3) Umm ..... VMWare?
4) Ohyeah. VMWare Server.

Oh, and *laugh*Accelerated-X*laugh*. Seriously. Who the heck uses that?

Oh, and maybe some random assorted browser plug-ins. Anything else? Anybody? Hello? ... didn't think so.

Seriously, why is this even worth Sun's time?! If I were a Sun shareholder (which I would never do, now that they have a "technology sharing" agreement with MS and are all buddy-buddy after accepting a settlement bribe from MS... well, I'd be frothing at the mouth even more than I am now. ;)

Re:WHAT Linux apps?

[MaoTse]

This well worth Sun PR's time.
To shut up all the people in here saying Sun is
not Open-Source enough ;-)

Re:WHAT Linux apps?

[Lukey Boy]

I use Accelerated-X on my Vaio notebook. It's the only thing I've found that supports XVideo, GL and most importantly the widescreen display. There's a bunch of VBIOS hacks that supposedly force the XFree/Xorg server to handle the resolution, but none of them have worked for me.

And Accelerated-X is available for Solaris also.

Actually...

[devphil]

I've used both Linux and Solaris for development for years. Was a sysadmin for both types of systems as well. And my dream operating system is something along the lines of GNU/Solaris.

Meaning it the same way that wackjob RMS means it: the GNU userspace utilities, with the Solaris kernel. I /really/ like some of the things that Solaris offers, but I vastly prefer the GNU command-line utils. Putting them together would make a nice, nice system.

Re:Actually...

[oldmanmtn]

Done (or at least getting there). Next time you're on a Solaris box, look in /usr/sfw/bin. Solaris now ships with bash (in /usr/bin) and GNU tar, grep, wget, texinfo, gs, ncftp (OK, not GNU but still usefull), and mozilla.

Re:Actually...

[wkcole]

The /usr/sfw tree is relatively new, having appeared in Solaris 9. Sun also ships (and has been for some years, at least since the 02/02 release of Sol8) an extra CD full of pre-built open source tools that generally install under /opt/sfw. They also sponsor sunfreeware.com, a site that maintains a staggering collection of pre-built freeware packages for Solaris versions back to 2.5.

Re:WOOOW

[Proteus]

so you can dump you free platform, buy an expensive one and run your free apps on them?

No, you can choose Solaris over a FOSS platform, get all the Solaris apps &c, and also run Linux apps. Given the choice between buying a decent Solaris machine for one application plus a decent Linux machine for another, or buying one really nice Solaris box that will do both jobs, I might just opt for the latter.

It's actually a very smart move by Sun - give people trying to decide between Solaris and Linux an easy out.

Re:In short, no.

[CoffeeDregs]

>blahblah...business model ... dead ... blah blah

I don't think so. There are still plenty of reasons to use closed-source software. But Moore's Law applies both to hardware and to software, so it's relatively easy to duplicate someone else's software a few years later. We have OS duplicates of Word and Excel, but are lagging on the dupe for Outlook (give me shared calendaring...).

What O/S does do for software is the same thing that the government does for motorists/business: provide a common infrastructure upon which to operate your car/app. I'd be happy to pay the government to develop my roads and a trucking company to get my stuff on those roads; and I'd be happy to hire some O/S specialists to develop, maintain and contribute to my O/S infrastructure and IBM to maintain my J2EE infrastructure (yes, yes, JBoss, but you get my point).

Business models involving incredibly expensive software and absurdly expensive support contracts will grow and thrive... But only in high-end areas which O/S and commodity hardware cannot address yet. This is common, though:

for (; 1 ;) {
low-end attacks high-end;
the high-end moves to the higher-end;
}

--- Drink the dregs

Then why do so many of us run it

[falcon5768]

I've considered them on the losing end of the battle until now.

Then why do a lot of major institutions still run Solaris??? We have a Win2000 network but still use a Sun box because for one of our packages, it ONLY works on a Sun box....

Sun isnt dead... its a nitch product like Apple is. For those sys admins who dont know better there is always Windows... but for people who actually care about getting something done right, and having up time measured in months or years there are the other OS's out there...

Will sun ever blow by Microsoft, no, but will it still be used? Well there is a story here about emulating 20 - 30 year old computer systems so you can bet your ass people will use it for years to come.

Apparently Apple is doing this too and for both I give congrats.... they are making it so Linux is more a flavor with no restrictions than right now where for some things you have to go with one or the other... A bold move on everyones part considering every one of these companies wants you to use THEIR version..

Re:IOW...

[Anonymous Coward]

Actually, many system calls aren't wrapped by glibc. To use those, you must #include and use the syscall macros, which use assembly, which includes the interrupt.

And statically-linked binaries DO count, because if your implementation can't run static binaries, it's not binary compatibility.

Besides.. The syscalls do not match one-to-one anyway. There are a lot of small incompatibilities between Linux and Sun syscalls. And then, Linux has Linux-only syscalls, Sun has Sun-only syscalls, etc.

Re:No

[sumbry]

FreeBSD, NetBSD, (and OpenBSD?) have been able to run Linux binaries natively for years and guess what - hardly anyone uses it.

Why? Because for the most part the same apps that can be run in binary mode is just as easily ported to the BSD anyways.

My guess is the same thing will happen with Solaris.

Re:typical for sun

[theProf]

Interesting thing about Sun's approach to 10 is that they are engineering a way to do without Linux at the application level.

If you get a nice 4-way opteron box, install solaris /x86, compartmentalise it via zones you get N effective linuxes (where N will be usefully large) on one chassis. Its gives you the same thing as IBM/VM/Linux, without buying a mainframe.

If you choose a server which has hardware supported by Solaris, then you don't actually need Linux at all. Oh, and Sun are pretty well imdenified already against patent risk with Solaris.

Re:IOW...

[oldmanmtn]

That's not insane, but it's probably not optimal either. :)

Hacking glibc to call the Solaris system calls (including all the necessary argument twiddling) would work in many cases, but it seems like a fragile solution:

Every time a glibc bug was found, Sun would have to reissue a patched glibc.

Statically compiled apps would never work.

System tools (like truss and dtrace) would report what glibc was doing instead of what the application was doing.

The only way I can see this working as a production-quality feature is if they actually implement the int80 mechanism and the Linux system calls in the Solaris kernel.

How do they handle the differences in threading models between Solaris ("real" threads) and Linux ("clone"ed processes that share address spaces)? Not a clue, but I'm sure it can't be done in a hacked up glibc.

SCO, Phase II

[Random BedHead Ed]

A few people above have already commented on this, but this is exactly what SCO did when it was Caldera. Caldera/SCO decided that since Linux was so popular, they could run Linux software on UNIX. They thought this was a pretty neat idea, even though no one was asking for it.

The flaw in this business plan is that if you have a lot of Linux apps, a good possible choice of OS to run them on is (and this is just a suggestion) Linux. There is rarely a need for a declining UNIX system to emulate a popular UNIX-like system. Perhaps if Caldera had grasped that its UNIX purchase was pointless and concentrated on out-Linuxing RedHat and SuSe, they'd still be called Caldera and would be in the Linux support and services business today. And hey, maybe they'd be earning a profit.

This is not to suggest that Solaris is a useless OS. It isn't. But if compatibility with Linux is something a customer wants, Linux is going to be the best choice, nine times out of ten. So the only possible reasons for Sun to do this are:

1. Lack of understanding of the OS market - not implausible.
2. Desire to be the go-to company if Linux becomes mired in lawsuit crossfire.

Re:SCO, Phase II

[swordgeek]

You're missing a key point here, I believe.

There are cases where people need Sun, and Sun apps. Lots of Geophysical apps run only on Solaris/Sparc right now. However, people might also want Linux apps, so making them available on the already mandatory Sun gear will keep some people gruntled.

Ultimately, you're right--if Linux compatibility is wanted, Linux is generally going to be the best solution in a vacuum. However if Linux compatibility is wanted on top of other requirements, then a compromise like this is better than having two machines on your desk.

Re:No

[Frit Mock]

If you don't have to care about your environment, you save time!

If your app just works as it is on another environment, more users are able to use it.

Of course you can reach the same goal with ./configure, but it does not realy matter how to reach that goal ... except perhaps, that a generalized soulution is preferable over an solution for an individual app.

#IFDEF SOLARIS is nice, but not a general solution, ./configure is nicer since it is more generalized, but come on compliance with another systems specification is the smoothest way to reach the goal of plattform-independency!

There is no better solution than this, to guarantee, that an application programmer can fully concentrate on his task and not fidle around with differences between different systems at any time! ./configure is more a workaround than a solution, but compliance to system-standards is a solution!

Re:OT but I want to say it anyway

[walt-sjc]

FWIW, that survey was done before any major adoption of linux by businesses (especially in a desktop role.) I would bet that if that survey were run again today and targeted the real customer base (business) instead of consumers / techies, the results would be different (although maybe not considering the lack of advancement of the product. Adobe seems to have killed it on all platforms.)

Pointing at a several year old survey as "proof" of viability for commercial apps on linux is silly. My company spends hundreds of thousands of dollars on commercial linux software every year. The linux market is alive and well for innovative products. Products that don't offer any real benefit over their free alternatives will not make it.

Frankly, the commercial word-processing / spreadsheet space is just about dead due to good free alternatives such as OpenOffice, Abiword / Gnumeric, Lyx, etc. Microsoft is going to find this out over the next couple years (I think office suites is their number 2 money maker IIRC...)

BTW, Commodore's horrible marketing killed Amiga and its market. Not piracy. I was an avid Amiga user / programmer from '85 to '95 when I switched to Linux.

Re:OT but I want to say it anyway

[Lussarn]

For the record, I hope you get moded to oblivion.

Re:OT but I want to say it anyway

[GreyWolf3000]

And this is one more reason why linux will never replace windows. No one would dare risk porting expensive commercial software to linux.

Unless, of course, free software can match commercial equivalents.

Re:No

[jonadab]

> the correct one would be to figure out your environment and build accordingly

No, the correct way is to build your software so that it makes no significant
difference what platform it's running on. With modern languages and libraries
and toolkits this is getting closer and closer to actually being possible.
(Think: Parrot and wxWindows.)

Re:No

[IvyKing]

One big difference between running Linux binaries on Solaris and Windoze binaries on OS/2 - Microsoft changed the Windoze ABI slightly after the agreement with IBM ran out - so that it was possible to write an application for Win 3.11 that would not run on OS/2. Since the Linux API and ABI is open, Sun is free to modify the Linux translation layer as necessary to maintain compatibility. Also remember that the whole reason for "Linux Standard Base" was to have a common ABI for LSB compliant distro's.

The benefit for Sun is that there are a significant number of Linux applications that do not come with source.

just a few days ago

[Knights who say 'INT]

Someone wondered if Linux would ever get certified as an UNIX.

Well, this is a true Unix getting certified as a Linux!

We are actually winning. Amazing.

Re:Apache

[pebs]

Are you kidding? That's not complicated? The only way it could be *more* complicated and convoluted is if you were also to tell me that I need to solder something onto my motherboard to get it to work. Compare this to ASP/IIS: either it comes already installed and running, or you end up putting a CD in the drive and clicking "next" a few times.

If you want easy, install Debian and type as root:

apt-get update; apt-get install apache2 mysql-server php4-mysql libapache2-mod-php4

Say yes to any prompts and you are done! Apache2/MySQL/PHP4 will now be setup and running.

Granted we were talking about installing on Solaris here, but your monkey ass had to bring up IIS like its some wonderfully simple thing that even licks your ass for you.

Now if there were a Debian Solaris that would be sweet.

Re:Apache

[NineNine]

a. FUD

b. FUD

C. FUD

D. FUD

E. You are incompetent.

I would've written a better response, but anyone with a half a brain already knows that you're totally full of shit.

Multiprocessor efficiency

[vlad_petric]

The reason Solaris does so well on many processors is because Sun's strategy regarding "big iron" was "ultraslows, but many". It's very true that their target workloads (server-based) can exploit multiprocessing easily. Having good support for multiprocessors was simply crucial for the platform.

In the x86 world things are quite different. Having been a desktop-oriented architecture for a long time, the main x86 chips (Opteron/Pentium IV) are pretty much the best these days at executing single-threaded stuff (see spec.org if you don't believe me). Multiprocessing was more of an "after-thought" than an initial requirement. Consequently, you can easily get 4-way SMPs for x86s, but not more than that (Sun AFAIK scales considerably better).

This reflects on x86 OSes as well. There's not that much need to do well on more than 8 execution contexts (4way SMP x2 - hyperthreading), and consequently having an operating system that scales better won't have that much of an impact on x86. Sure, in the "big iron" category things will be different, but not for the dominant architecture

Re:Multiprocessor efficiency

[T-Ranger]

Linuxs initial target was desktops and smaller systems. Conventional wisdom is that Intel procs before the Pentium Pro could not do SMP, but that is an over simplification. They were the first that could do SMP easily. At least as far back as 1981 people were making "Supercomputers" out of Intel CPUs. http://www.computerhistory.org/VirtualVisibleStora ge/artifact_main.php?tax_id=03.04.07.00

Linus thinks working on small, embedded, systems is perhaps more important then massive scaleability with SMP. That said: SGI has been producing massivly SMPd machines that run Linux for a long time. Prob with MIPS first, definitly with IA32 now, and with a heavily hacked kernel, but still Linux.

I think Suns strategy of "ultraslows, but many" is much older then their "big iron" history. At least as far back as 32bit Sparcs, desktops, not to mention "workgroup" servers, were SMP capable. I dont know if it is realy a strategy so much as a necessity given RISC style processors. Of course, extending that argument, IA32/64 systems, compared to 370 arch IBM Mainframes are RISC too. And to scale up IA systems, you go SMP. Comparing Sun and Intel micro computers to a mainframe might be compleatly useless though :)

Re:Apache

[linsys]

Your right... I am full of it, don't take my input read it from ORACLE!

http://www.oracle.com/technology/pub/articles/hull _asp.html

Speed :
PHP STRONG
ASP WEAK


Platform:
PHP STRONG
ASP WEAK


Platform Price:
PHP FEEE
ASP $$$


Security and I quote "Security. ASP.NET runs on IIS, which has been compromised innumerable times, as evidenced by IT news reports every other week. It has become such a liability, in fact, that in spite of all the marketing dollars spent on it, many IT professionals refuse to have their networks exposed with an IIS Web server. PHP, however, works with Apache, which has a proven track record of speed, reliability, and hardened security. Check www.securityfocus.com for more information."

Thanks for playing... I understand you probably spent years learning IIS/ASP/Win2K and now you realize it sux so you spend your time bashing real operating systems and REAL programming languages so you don't feel like your years at ITT TECH where a waste of time...

Re:OT but I want to say it anyway

[huchida]

Piracy killed commercial software on the Amiga, and I think the same would happen on Linux if things like Office, Photoshop and a lot of games were ever ported to it.

"Pirating" has never killed a system. People have been "pirating" apps for every computer from the TRS-80 to Windows XP. If anything, access to "free" (albeit less-than-legally acquired) programs and games is a pretty strong selling point... If one has all the Windows apps they need, if a friend says he can get them any Windows program, if they know where to get them on my own, then they're less likely to switch to a Mac with my next system. Not saying it's right, but it's rampant, even-or especially- on the most widely used OS'es. I've seen churches using copies of Photoshop and Office that were obviously pirated without a second thought.

In any event, Linux users may not want to pay for commercial software, and may even vocally bellyache about it... But it's not users who really matter. Businesses will pay, because they have reasons for switching to Linux beyond getting stuff for free.

Re:Apache

[NineNine]

No, kid. I have experience. The products are virtually the same. They do the same ting, and one is as secure as the other. In fact, studies have shown that IIS is much faster than Apache in circumstances. Plus, ASP doesn't cost a dime. It comes with W2K Server.

And next time you quote some supposed article, try to at least point to a real article, and not a 404.

Re:Apache

[linsys]

Slashdot is putting in an extra . In the comment field it works then I preview it and a space gets put in between hull and the _ (that's an underscore).

http://www.oracle.com/technology/pub/articles/hu ll _asp.html

I wouldn't expect a NT admin to figure that out, it involved troubleshooting...

Linux and "no devel branch"

[0x0d0a]

And there are way too many patches going into an allegedly "stable" tree.

This your evaluation based on reading LKML, or secondhand?

Because, honestly, Slashdot did an article on this a bit back and it was prety negative about it, and I don't really agree with the article author's opinions there.

I personally don't see it as a problem. What most users (business and whatnot) consider "stable" is a kernel that's actually been tested with their userspace software, which means a kernel from a distro maker. Linus is supposedly adding smaller features now. With the introduction of the "RC" version numbering scheme, he's basically making a small development branch for each new stable release. I'd say that this is more about reducing the size of branches in kernels being fed to the distro manufacturers (which really isn't what you should be using anyway, if you're running a production machine) from year-long-plus beasts into two-month-long branches or so (the current "devel branch" is 2.6.8-rc3, which has been branched from 2.6.7 for about two months).

Re:Linux and "no devel branch"

[Cajal]

Actually, it's based on reading kerneltrap and on my own experiences. There are still fairly large changes going into the "stable" tree (check out LWN's kernel page each week). And why shouldn't you be using a vanilla kernel on a production machine?

Re:Linux and "no devel branch"

[0x0d0a]

And why shouldn't you be using a vanilla kernel on a production machine?

Because while the kernel has been tested to some extent, is hasn't been tried with your mix of userspace software. It'd be like taking the latest kernel that passes the internal Microsoft QA and immediately slapping it into, say, NT4, without internal or beta testing of the whole at all.

Re:Linux and "no devel branch"

[Cajal]

But if the kernel is properly developed, then it shouldn't matter. Besides, most user-space apps shouldn't be making calls that could vary that much from one kernel to another.

This is the issue they're already facing

[Xtifr]

One word: Oracle. The database giant says, "Sun is our primary and recommended platform," and Sun enjoys a decade of dominance in the server market. Then Oracle says, "Linux is now our primary and recommended platform," and suddenly Sun is struggling to make ends meet. Coincidence? I think not.

Go to www.oracle.com and click on "technologies". What do you see? You see Linux (and, to be fair, Windows). What don't you see? You don't see Solaris. Hmmm....

Re:No

[anothy]

...the correct one would be to figure out your environment and build accordingly... ...thanks to the wonders of "./configure ; make"... ...it just gets compiled and run...

yet, at the same time:

...it's about those developers becoming less sloppy and making fewer assumptions about environment.

configure and autoconf is not the way to write "portable" code, and certainly not the way to "figure out your environment". configure is a hideous hack, built to accommodate the vast array of non-portable code and busted interfaces people would like to code to, and encourages people to write further non-portable code that only looks portable because configure covers your back. write to defined interfaces, use portable libraries, and there's no need for configure.

the biggest argument for configure is that it allows you to work around broken interfaces supplied by the vendor, particularly in systems where you don't have source. but the amount of work that's been poured into configure and autoconf to try to deal with these bugs in a pseudo-portable, pseudo-transparent manner is dramatically greater than the effort it takes to write a library that conforms to the portable interface. that's a much cleaner way to go, both for the code you're delivering (less code to accommodate exceptions) and the interfaces in the long term (less code weight to continue to support interface bugs). ever tried porting configure to a new platform? ugh.

Re:IOW...

[jonabbey]

Yeah, but you can't do that between all versions of the Linux kernel.. see an NPTL kernel versus one without..

I take the point about statically bound apps and the complexity of supporting a custom glibc fork, but I'm sure there'll be provisios and limits on what versions of the Linux ABI are supported with this setup.

Yeah...

[sp0rk173]

Yup. On my BSD machines I only use linux binaries for a few applications - flash is the only thing that comes to mind. Actually, the first thing that came to mind when i read that was FreeBSD's linuxulator. I wonder if sun used any code from FreeBSD's solution or not. Not that there's anything wrong with that...just a curiosity.

Linux? Run?

[sp0rk173]

Pfftt. Linux doesn't run on anything - it doesn't have legs. It's a kernel. Have you ever seen a popcorn kernel run? I, for one, have not.

This will shut SCO up

[ajs318]

As soon as there is an Open Source version of System V, that's the end of the line for SCO. The next Linux kernel release following the Open Sourcing of Solaris could -- entirely legally -- have been based on Open Source Solaris.

Mind, SCO's tail seems firmly between their legs right now -- IMHO they hope to disappear quietly into the night, and maybe nobody will come chasing after them to finish them off.

Re:Apache

[Zebbers]

I'm spoiled.....I just emerge ;)

Re:No

[Syberghost]

The question was stupid anyway.

Nobody buys Linux instead of Solaris because it's got cool apps.

We buy Linux instead of Solaris because we don't want to pay $8 bajillion for a system that doesn't need to be that robust; where the application can run in parallel on lots of cheap hardware and you don't care if one or two or fifty of them die, Sun makes no sense.

Once again, Sun doesn't get it.

Re:No

[Lodragandraoidh]

...and the vendors should know better...

Hahahahahahaha!!!! You gotta be kidding me.

Re:OT but I want to say it anyway

[chegosaurus]

I'd give you the mod points, but I doubt you could work out how to use them.

No they would not!

[I_am_the_man]

http://www.sun.com/third-party/global/oracle/Sun_O racle_DS_final.pdf

Re:This may be a new SUNrise.... or maybe not

[sad_]

1. Linux apps will run on Solaris 10 on Intel/Sparc. Someone said this is just for X86.
it is on the website of sun as well. it works only on Solaris x86!

I'm not convinced

[sad_]

I'm not sure why i would want to use this feature. it only works for x86, as you can read here.

although solaris x86 is not locked to specific hardware, I doubt SUN will provide your company support on non-SUN hardware. So you will be running Solaris x86 on SUN hardware which will be running linux apps, while I could just as well run native linux on x86 on probably cheaper hardware. Tell me exactly why should prefer running linux apps under Solaris instead? Certainly on x86 this makes no sense.

What this is, is just a way for SUN to say that Solaris x86 is a vailable platform, because it runs all these applications (which are linux apps). Now they no longer have to worry about pushing vendors to port apps to Solaris x86.

However I don't see this as a winner for SUN, apps will still not be written for Solaris x86 and how long will they be able to keep this game up? What if apps start using specific kernel calls which solaris does not have. I'd be surprised if a command like 'iptables' would work, while it is a very basic linux command.

Re:No

[addaon]

Unless your software is one of those 'modern languages and libraries and toolkits'.

Re:OT but I want to say it anyway

[black mariah]

Too bad everything you linked to is a half-assed implementation that is matched in quality by most $20 Windows shareware apps. I don't understand why you use both Gimp and Cinepaint, as they're basically the same thing. Are you that desperate for examples of good free software alternatives? Ardour is a fucking joke, as is JACK. Blender is simply nowhere near being a quality 3d app, and unless the developers pull their heads out of their ass and push a concerted effort to make it less intentionally fucking weird then it will always be a half-assed program.

Re:OT but I want to say it anyway

[GreyWolf3000]

Gimp and Cinepaint (despite originally being a fork of Gimp) are not the same thing, Ardour is not a joke, and have you seen the screenshots of what people have done with Blender?

I'll accept what you're saying for a second. Suppose they do suck. That doesn't run contrary to my point, since I suggested that free software alternatives may reach the level of the commercial apps.

My point was that your point that Linux will never replace Windows because users expect software to be free, thus barring any commercial software from being ported.

That is a silly point to begin with, since most people don't use that software, and if 75% of the home desktop users who don't use all that software used Linux, the commercial vendors would have to port to Linux (and at that point, many, many users wouldn't expect their software to be free).

Can you please tell me why JACK and Ardour suck? As well as Gimp and Cinepaint. Also, your beef against Blender appears to be the UI; note that many experienced 3d artists grow to love the UI.

At any rate, I see no architectural stumbling blocks.

can they become stronger

[jdkane]

Will the power of Linux apps put Solaris back into the running?

The power of more apps on the platform can certainly not hurt Sun. However why not run Linux apps on Linux instead. I've read Solaris is very stable and has some enterprise features that Linux may not support (maybe that's on the hardware side). However isn't Linux as stable as Solaris? If Solaris will become OSS, how about merging the best of the two. Does this help the potential copyright/patent problems that Linux may face -- e.g. the city of Munich researching possible infringements which is slowing down adoption of Linux?

Re:OT but I want to say it anyway

[black mariah]

Gimp and Cinepaint are the exact same damn thing, save for about two features. My issue with Blender is less about the interface (now. They're made major improvements) and more with how everything works completely different from every other app ever made, for no good reason whatsoever. Yes, I've seen some of the stuff made with Blender but I've also seen FAR better stuff made with Truespace, which is an even bigger POS. Good artists are good artists, regardless of software.

I've detailed my displeasure with Ardour and JACK in one of my journals. Check it out sometime.

Embrace and extend?

[RhettLivingston]

Could it be that Sun is learning from Microsoft? This means that Solaris will run Linux stuff, but not that all Solaris code will run on Linux. I'd bet whatever OSI license they choose won't allow many Linux distributions to adopt much of their code. Seems like this is all a recipe for lots of problems unless everyone converts to Solaris. I'd bet that's there hope. Then their market for selling extensions above and beyond the open source version would be larger.

color scheme (OT)

[gargan]

Of all the god-awful color schemes the /. editors come up with, this one I like. I have also never noticed it before. I'm thinking of installing proxomitron or something to change all the slashdot urls to linux.slashdot.org as I recently discovered you could do that to change the color scheme.

anyway, carry on.

Re:No

[IceFox]

Yes, five minute after making the post I thought to add a little anendem that it really only applies to little home grown type apps (but away from the internet the rest of the day). Stuff that you are selling to primarily run on solaris boxes anyway is different, in that case the reason you are primarily on Solaris isn't why would would distribute Linux binaries (hw stability etc).

-Benjamin Meyer

What about the compiler?

[vegetasaiyajin]

How is one supposed to compile the open source solaris?
It is very well known that gcc generates slow code for Sparc, and I don't even know if it can compile it at all.
They should at least provide a free beer version of their compiler.
I have a sparc server at work and it pisses me off that they don't provide a C compiler.

VMware

[Mistah Blue]

So, does this mean you could run VMware with Solaris x86 as the host os? That would be interesting for me.

....funny....

[Zet2002]

funny...i get the slashdot newsletters...and im starting to see the word "linux" alot. Especially with subjects that state that linux is uncluded on something. I like it.

Re:No

[JonAnderson]

Hmm, If you buying a Linux with support (i.e. Redhat AS) and running it on X86 then it is a myth that Sun is more expensive. That's simply not the case anymore. Sun is VERY competetive in price in the upto 4 way X86 market and still has good TCO when looking at the larger sparc servers. In the markets where Sun deals in the most Linux isn't free.....

Re:Linux? Run?

[sp0rk173]

communist.

How to make sense of Sun's attitude ?

[crashedbutterfly]

These statements have been made over the last two months at large :

Hardware is a commodity, it will be free :
http://it.slashdot.org/article.pl?sid=04/06/01/162 7236&tid=102&tid=137/

Solaris will be OSS
http://it.slashdot.org/article.pl?sid=04/06/02/133 256&tid=102&tid=190&tid=130/
http://linux.slashdot.org/article.pl?sid=04/08/04/ 1248251&tid=102&tid=163&tid=130/

Java will or will not be OSS
Will : http://developers.slashdot.org/article.pl?sid=04/0 6/04/002224&tid=108&tid=156&tid=102&tid=8/
Won't: http://developers.slashdot.org/article.pl?sid=04/0 6/05/183209&tid=108&tid=156&tid=102&tid=8/

Leaves us with... A basically free OS, running on a free hardware able to run free software out of the box.
The only business model they seem to aim for is maintenance and services...

But... An OS as polished as Sun becoming free (and already available on x86) might become a new contender to Windows desktop/workstation. For this move they lack applications, but making it able to run Linux apps out of the box seems a smart move to close the gap...

Any feeling on those matters ?